<?xml version="1.0"encoding="iso-8859-1"?> <?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?> <?rfc strict="no" ?> <?rfc toc="yes"?> <?rfc tocompact="yes"?> <?rfc tocdepth="3"?> <?rfc tocindent="yes"?> <?rfc symrefs="yes"?> <?rfc sortrefs="yes" ?> <?rfc comments="yes"?> <?rfc inline="yes"?> <?rfc compact="yes"?> <?rfc subcompact="no"?>encoding="UTF-8"?> <!-- updated by Chris 09/15/20 --> <!DOCTYPE rfc SYSTEM"rfc2629.dtd" [ <!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml"> <!ENTITY RFC3022 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3022.xml"> <!ENTITY RFC4271 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4271.xml"> <!ENTITY RFC4272 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4272.xml"> <!ENTITY RFC4360 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4360.xml"> <!ENTITY RFC4364 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4364.xml"> <!ENTITY RFC4760 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4760.xml"> <!ENTITY RFC5925 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5925.xml"> <!ENTITY RFC6146 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6146.xml"> <!ENTITY RFC6296 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6296.xml"> <!ENTITY RFC6952 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6952.xml"> <!ENTITY RFC7432 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7432.xml"> <!ENTITY RFC7498 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7498.xml"> <!ENTITY RFC7606 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7606.xml"> <!ENTITY RFC7665 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7665.xml"> <!ENTITY RFC8126 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8126.xml"> <!ENTITY RFC8174 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8174.xml"> <!ENTITY RFC8300 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8300.xml"> <!ENTITY RFC8595 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8595.xml"> <!ENTITY RFC8596 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8596.xml"> ]>"rfc2629-xhtml.ent"> <rfccategory="std"xmlns:xi="http://www.w3.org/2001/XInclude" docName="draft-ietf-bess-nsh-bgp-control-plane-18"ipr="trust200902">number="9015" ipr="trust200902" obsoletes="" updates="" submissionType="IETF" category="std" consensus="true" xml:lang="en" tocInclude="true" tocDepth="3" symRefs="true" sortRefs="true" version="3"> <!-- xml2rfc v2v3 conversion 3.1.1 --> <front> <title abbrev="BGP for NSH SFC">BGP Control Plane for the Network Service Header in Service Function Chaining</title> <seriesInfo name="RFC" value="9015"/> <author fullname="Adrian Farrel" initials="A." surname="Farrel"> <organization>Old Dog Consulting</organization> <address> <email>adrian@olddog.co.uk</email> </address> </author> <author fullname="John Drake" initials="J." surname="Drake"> <organization>Juniper Networks</organization> <address> <email>jdrake@juniper.net</email> </address> </author> <author fullname="Eric Rosen" initials="E." surname="Rosen"> <organization>Juniper Networks</organization> <address> <email>erosen52@gmail.com</email> </address> </author> <author fullname="Jim Uttaro" initials="J." surname="Uttaro"> <organization>AT&T</organization> <address> <email>ju1738@att.com</email> </address> </author> <author fullname="Luay Jalil" initials="L" surname="Jalil"> <organization>Verizon</organization> <address> <email>luay.jalil@verizon.com</email> </address> </author> <dateyear="2020"year="2021" month="June" /> <area>Routing</area> <workgroup>BESS Working Group</workgroup> <keyword>Service Function Chaining</keyword> <keyword>Service Function Chain</keyword> <keyword>Network Service Header</keyword> <keyword>Service Function</keyword> <keyword>Service Function Forwarder</keyword> <keyword>Service Function Path</keyword> <keyword>Service Function Path Route</keyword> <keyword>Service Function Instance</keyword> <keyword>Service Function Instance Route</keyword> <keyword>Service Function Type</keyword> <keyword>Control Plane</keyword> <abstract> <t>This document describes the use of BGP as a control plane for networks that supportService Function Chaining (SFC).service function chaining. The document introduces a new BGP address family called theSFC"Service Function Chain (SFC) Address Family Identifier / Subsequent Address FamilyIdentifierIdentifier" (SFC AFI/SAFI) with tworoute types.Route Types. Oneroute typeRoute Type is originated by a node to advertise that it hosts a particular instance of a specified service function. Thisroute typeRoute Type also provides "instructions" on how to send a packet to the hosting node in a way that indicates that the service function has to be applied to the packet. The otherroute typeRoute Type is used by aControllercontroller to advertise the paths of "chains" of servicefunctions,functions andtogive a unique designator to each such path so that they can be used in conjunction with the Network Service Header (NSH) defined in RFC 8300.</t> <t>This document adopts theSFCservice function chaining architecture described in RFC 7665.</t> </abstract> </front> <middle> <section anchor="introduction"title="Introduction">numbered="true" toc="default"> <name>Introduction</name> <t>As described in <xref target="RFC7498"/>,format="default"/>, the delivery of end-to-end services can require a packet to pass through a series of Service Functions (SFs)(e.g.,-- e.g., WAN and application accelerators, Deep Packet Inspection (DPI) engines, firewalls, TCP optimizers, and server loadbalancers)balancers -- in a specifiedorder:order; this is termed"Service Function Chaining" (SFC)."service function chaining". There are a number of issues associated with deploying and maintaining service function chaining in production networks, which are described below.</t> <t>Historically, if a packet needed to travel through a particular service chain, the nodes hosting the service functions of that chain were placed in the network topology in such a way that the packet could not reach its ultimate destination without first passing through all the service functions in the proper order. This need to place the service functions at particular topological locations limited the ability to adapt a service function chain to changes in network topology (e.g., link or node failures), network utilization, or offered service load. These topological restrictions on where the service functionscancould be placed raised the following issues:<list style="numbers"> <t>The</t> <ol spacing="normal" type="1"><li>The process of configuring or modifying a service function chain is operationally complex and may require changes to the networktopology.</t> <t>Alternatetopology.</li> <li>Alternate or redundant service functions may need to be co-located with the primary servicefunctions.</t> <t>Whenfunctions.</li> <li>When there is more than one path between source and destination, forwarding may beasymmetricasymmetric, and it may be difficult to support bidirectional service function chains using simple routing methodologies and protocols without adding mechanisms for traffic steering or trafficengineering.</t> </list></t>engineering.</li> </ol> <t>In order to address these issues, theSFCservice function chaining architecture describesService Function Chainsservice function chains that are built in their own overlay network (the service function overlay network), coexisting with other overlay networks, over a common underlay network <xref target="RFC7665"/>.format="default"/>. AService Function Chainservice function chain is a sequence ofService Functionsservice functions through which packet flows that satisfy specified criteria will pass.</t> <t>This document describes the use of BGP as a control plane for networks that supportService Function Chaining (SFC).service function chaining. The document introduces a new BGP address family called theSFC"Service Function Chain (SFC) Address Family Identifier / Subsequent Address FamilyIdentifier (AFI/SAFI)Identifier" (SFC AFI/SAFI) with tworoute types.Route Types. Oneroute typeRoute Type is originated by a node to advertise that it hosts a particular instance of a specified service function. Thisroute typeRoute Type also provides "instructions" on how to send a packet to the hosting node in a way that indicates that the service function has to be applied to the packet. The otherroute typeRoute Type is used by aControllercontroller (a centralized network component responsible for planning and coordinatingService Function Chainingservice function chaining within the network) to advertise the paths of "chains" of servicefunctions,functions andtogive a unique designator to each such path so that they can be used in conjunction with the Network Service Header (NSH) <xreftarget="RFC8300"/>.</t>target="RFC8300" format="default"/>.</t> <t>This document adopts theSFCservice function chaining architecture described in <xreftarget="RFC7665"/>.</t>target="RFC7665" format="default"/>.</t> <sectiontitle="Requirements Language">numbered="true" toc="default"> <name>Requirements Language</name> <t>The key words"MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY","<bcp14>MUST</bcp14>", "<bcp14>MUST NOT</bcp14>", "<bcp14>REQUIRED</bcp14>", "<bcp14>SHALL</bcp14>", "<bcp14>SHALL NOT</bcp14>", "<bcp14>SHOULD</bcp14>", "<bcp14>SHOULD NOT</bcp14>", "<bcp14>RECOMMENDED</bcp14>", "<bcp14>NOT RECOMMENDED</bcp14>", "<bcp14>MAY</bcp14>", and"OPTIONAL""<bcp14>OPTIONAL</bcp14>" in this document are to be interpreted as described in BCP 14 <xreftarget="RFC2119"/>target="RFC2119" format="default"/> <xreftarget="RFC8174"/>target="RFC8174" format="default"/> when, and only when, they appear in all capitals, as shown here.</t> </section> <section anchor="terms"title="Terminology">numbered="true" toc="default"> <name>Terminology</name> <t>This document uses the following terms from <xreftarget="RFC7665"/>: <list style="symbols"> <t>Bidirectionaltarget="RFC7665" format="default"/>: </t> <ul spacing="normal"> <li>Bidirectional Service FunctionChain</t> <t>Classifier</t> <t>ServiceChain</li> <li>Classifier</li> <li>Service Function(SF)</t> <t>Service(SF)</li> <li>Service Function Chain(SFC)</t> <t>Service(SFC)</li> <li>Service Function Forwarder(SFF)</t> <t>Service(SFF)</li> <li>Service Function Instance(SFI)</t> <t>Service(SFI)</li> <li>Service Function Path(SFP)</t> <t>SFC branching</t> </list></t>(SFP)</li> <li>SFC branching</li> </ul> <t>Additionally, this document uses the following terms from <xreftarget="RFC8300"/>: <list style="symbols"> <t>Networktarget="RFC8300" format="default"/>: </t> <ul spacing="normal"> <li>Network Service Header(NSH)</t> <t>Service(NSH)</li> <li>Service Index(SI)</t> <t>Service(SI)</li> <li>Service Path Identifier(SPI)</t> </list></t>(SPI)</li> </ul> <t>This document introduces the following terms:<list style="symbols"> <t>Service</t> <dl> <dt>Service Function Instance Route(SFIR). A(SFIR):</dt><dd>A new BGP Route Type advertised by the node that hosts an SFI to describe the SFI and to announce the way to forward a packet to the node through the underlaynetwork.</t> <t>Servicenetwork.</dd> <dt>Service Function OverlayNetwork. TheNetwork:</dt><dd>The logical network comprised ofClassifiers,classifiers, SFFs, and SFIs that are connected by paths or tunnels through underlay transportnetworks.</t> <t>Servicenetworks.</dd> <dt>Service Function Path Route(SFPR). A(SFPR):</dt><dd>A new BGP Route Type originated byControllerscontrollers to advertise the details of eachSFP.</t> <t>ServiceSFP.</dd> <dt>Service Function Type(SFT). An(SFT):</dt><dd>An indication of the function and features of anSFI.</t> </list></t>SFI.</dd> </dl> </section> </section> <section anchor="overview"title="Overview">numbered="true" toc="default"> <name>Overview</name> <t>This section provides an overview ofService Function Chainingservice function chaining ingeneral,general and the control plane defined in this document. After reading this section, readers may find it helpful to look through <xref target="example"/>format="default"/> for some simple worked examples.</t> <section anchor="funcover"title="Overviewnumbered="true" toc="default"> <name>Overview of Service FunctionChaining">Chaining</name> <t>In <xreftarget="RFC8300"/>target="RFC8300" format="default"/>, a Service Function Chain (SFC) is an ordered list of Service Functions (SFs). A Service Function Path (SFP) is an indication of which instances of SFs are acceptable to be traversed in an instantiation of an SFC in a service function overlay network. The Service Path Identifier (SPI) is a 24-bit number that identifies a specific SFP, and a Service Index (SI) is an 8-bit number that identifies a specific point in that path. In the context of a particular SFP (identified by an SPI), an SI represents a particularService Function,service function and indicates the order of that SF in the SFP.</t> <t>Within the context of a specific SFP, an SI references a set of one or more SFs. Each of those SFs may be supported by one or more Service Function Instances (SFIs).ThusThus, an SI may represent a choice of SFIs of one or moreService Function Types.service function types. By deploying multiple SFIs for a single SF, one can provide load balancing and redundancy.</t> <t>A special functional element, called aClassifier,"classifier", is located at each ingress point to a service function overlay network. It assigns the packets of a given packet flow to a specificService Function Path.SFP. This may be done by comparing specific fields in apacket'spacket's header with local policy, which may be customer/network/service specific. TheClassifierclassifier picks an SFP and sets the SPIaccordingly,accordingly; it then sets the SI to the value of the SI for the first hop in the SFP, and then prepends a NetworkServicesService Header (NSH) <xref target="RFC8300"/>format="default"/> containing the assigned SPI/SI to that packet. Note that theClassifierclassifier and the node that hosts the firstService FunctionSF ina Service Function Pathan SFP need not be located at the same point in the service function overlay network.</t> <t>Note that the presence of the NSH can make it difficult for nodes in the underlay network to locate the fields in the original packet that would normally be used to constrainequal costequal-cost multipath (ECMP) forwarding. Therefore, it is recommended that the node prepending the NSH also provide some form of entropy indicator that can be used in the underlay network. How this indicator is generated and supplied, and how an SFF generates a new entropy indicator when it forwards a packet to the next SFF, are out of the scope of this document.</t> <t>The Service Function Forwarder (SFF) receives a packet from the previous node ina Service Function Path,an SFP, removes thepacket'spacket's link layer or tunnelencapsulationencapsulation, and hands the packet and the NSH to theService Function InstanceSFI for processing. The SFI has no knowledge of the SFP.</t> <t>When the SFF receives the packet and the NSH back from theSFISFI, it must select the next SFI along the path using the SPI and SI in the NSH and potentially choosing between multiple SFIs (possibly of differentService Function Types)SFTs), as described in <xref target="selection"/>.format="default"/>. In the normalcasecase, the SPI remainsunchangedunchanged, and the SI will have been decremented to indicate the next SF along the path. But other possibilities exist if the SF makes other changes to the NSH through a process ofre-classification: <list style="symbols">reclassification: </t> <ul spacing="normal"> <li> <t>The SI in the NSH may indicate:<list style="symbols"> <t>A</t> <ul spacing="normal"> <li>A previous SF in thepath:path; this is known as "looping" (see <xreftarget="looping"/>).</t> <t>Antarget="looping" format="default"/>).</li> <li>An SF further down thepath:path; this is known as "jumping"(see also(again see <xreftarget="looping"/>).</t> </list></t> <t>Thetarget="looping" format="default"/>).</li> </ul> </li> <li>The SPI and the SI may point to an SF on a differentSFP:SFP; this is known as "branching" (seealso<xreftarget="looping"/>).</t> </list></t>target="looping" format="default"/>).</li> </ul> <t>Such modifications are limited to within the same service function overlay network. That is, an SPI is known within the scope of service function overlay network. Furthermore, the new SI value is interpreted in the context of the SFP identified by the SPI.</t> <t>As described in <xreftarget="RFC8300"/>,target="RFC8300" format="default"/>, an SPI that is unknown orinvalid SPInot valid is treated as anerrorerror, and the SFF drops thepacket:packet; such errors should be logged, and such logs are subject to rate limits.</t> <t>Also, as described in <xref target="RFC8300"/>,format="default"/>, an SFF receiving an SI that is unknown in the context of the SPI can reduce the value to the next meaningful SI value in the SFP indicated by the SPI. If no such valueexistsexists, or if the SFF does not support reducing the SI, the SFF drops the packet and should log theevent:event; such logs are also subject to rate limits.</t> <t>The SFF then selects an SFI that provides the SF denoted by theSPI/SI,SPI/SI and forwards the packet to the SFF that supports that SFI.</t> <t><xref target="RFC8300"/>format="default"/> makes it clear that the intended scope is for use within a singleprovider'sprovider's operational domain.</t> <t>This document adopts theSFCservice function chaining architecture described in <xref target="RFC7665"/>format="default"/> and adds a control plane to support thefunctionsfunctions, as described in <xref target="ctrlover"/>.format="default"/>. An essential component of this solution is theController.controller. This is a network component responsible for planning SFPs within the network. It gathers information about the availability of SFIs and SFFs, instructs the control plane about the SFPs to be programmed, and instructs theClassifiersclassifiers how to assign traffic flows to individual SFPs.</t> </section> <section anchor="ctrlover"title="Controlnumbered="true" toc="default"> <name>Control PlaneOverview">Overview</name> <t>To accomplish the function described in <xref target="funcover"/>,format="default"/>, this document introduces the Service Function Type(SFT) that(SFT), which is the category of SF that is supported by an SFF (such as "firewall"). An IANA registry ofService Function Typesservice function types is introduced in <xref target="SFTreg"/>format="default"/> and is consistent with types used in otherworkwork, such as <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments"/>.format="default"/>. An SFF may support SFs of multiple different SFTs, and it may support multiple SFIs of each SF.</t> <t>The registry of SFT values (seeSection 10.5)<xref target="SFTreg"/>) is split into three ranges with assignment policies per <xref target="RFC8126"/>: <list style="symbols"> <t>The Special Purposeformat="default"/>: </t> <ul spacing="normal"> <li>The special-purpose SFT values range is assigned through Standards Action. Values in that range are used for special SFC operations and do not apply to the types of SF that may form part of theSFP.</t> <t>TheSFP.</li> <li>The First Come First Served range tracks assignments ofSTFSFT values made by any party that defines an SF type. Reference through an Internet-Draft is desirable, but notrequired.</t> <t>Therequired.</li> <li>The Private Use range is not tracked by IANA and is primarily intended for use in private networks where the meaning of the SFT values is locally tracked and under the control of a localadministrator.</t> </list></t>administrator.</li> </ul> <t>It is envisaged that the majority of SFT values used will be assigned from the First Come First Served space in the registry. This will ensureinteroperabilityinteroperability, especially in situations where software and hardware from different vendorsisare deployed in the same networks, or when networks are merged. However, operators of private networks may choose to develop their own SFs and manage the configuration and operation of their network through their own list of SFT values.</t> <t>This document also introduces a new BGP AFI/SAFI (valuesto be assigned by IANA)31 and 9, respectively) for "SFC Routes". Two SFC Route Types are defined by this document: the Service Function Instance Route(SFIR),(SFIR) and the Service Function Path Route (SFPR). As detailed in <xref target="sfcBGPRoutes"/>,format="default"/>, theroute typeRoute Type is indicated by asub-fieldsubfield in the Network Layer Reachability Information (NLRI).<list style="symbols"> <t>The</t> <ul spacing="normal"> <li>The SFIR is advertised by the node that provides access to the service function instance (i.e., the SFF). The SFIR describes a particular instance of a particularService FunctionSF (i.e., an SFI) and the way to forward a packet to it through the underlay network, i.e., IP address and encapsulationinformation.</t>information.</li> <li> <t>The SFPRs are originated byControllers.controllers. One SFPR is originated for eachService Function Path.SFP. The SFPR specifies:<list style="letters"> <t>the</t> <ol spacing="normal" type="A"><li>the SPI of thepath</t> <t>thepath,</li> <li>the sequence of SFTs and/or SFIs of which the pathconsists</t> <t>forconsists, and</li> <li>for each such SFT or SFI, the SI that represents it in the identifiedpath.</t> </list></t> </list></t>path.</li> </ol> </li> </ul> <t>This approach assumes that there is an underlay network that provides connectivity between SFFs andControllers,controllers and that the SFFs are grouped to form one or more service function overlay networks through which SFPs are built. We assume that theControllerscontrollers have BGP connectivity to all SFFs and allClassifiersclassifiers within each service function overlay network.</t> <t>When choosing the next SFI in a path, the SFF uses the SPI and SI as well as the SFT to choose among the SFIs, applying, for example, aload balancingload-balancing algorithm or direct knowledge of the underlay networktopologytopology, as described in <xref target="mode"/>.</t>format="default"/>.</t> <t>The SFF then encapsulates the packet using the encapsulation specified by the SFIR of the selected SFI and forwards the packet. See <xref target="SFCarch"/>.</t> <t>Thusformat="default"/>.</t> <t>Thus, the SFF can be seen as a portal in the underlay network through which a particular SFI is reached.</t> <t><xreftarget="SFCarch"/>target="SFCarch" format="default"/> shows a reference model for theSFCservice function chaining architecture. There are four SFFs (SFF-1 through SFF-4) connected by tunnels across the underlay network. Packets arrive at aClassifierclassifier and are channeled along SFPs to destinations reachable through SFF-4.</t> <t>SFF-1 and SFF-4 each have one instance of one SF attached (SFa and SFe). SFF-2 has two types of SF attached:there isone instance of one(SFc),(SFc) and three instances of the other (SFb). SFF-3 has just one instance of an SF (SFd), butitin thiscasecase, the type of SFd is the same type as SFb (SFTx).</t> <t>This figure demonstrates how load balancing can be achieved by creating several SFPs that satisfy the same SFC. Suppose an SFC needs to include SFa, an SF of type SFTx, and SFc. A number of SFPs can be constructed using any instance of SFb or using SFd. Load balancing may be applied at two places:<list style="symbols"> <t>The Classifier</t> <ul spacing="normal"> <li>The classifier may distribute different flows onto different SFPs to share the load in the network and acrossSFIs.</t> <t>SFF-2SFIs.</li> <li>SFF-2 may distribute different flows (on the same SFP) to different instances of SFb to share the processingload.</t> </list></t>load.</li> </ul> <t>Note that, for convenience and clarity, <xref target="SFCarch"/>format="default"/> shows only a few tunnels between SFFs. There could be a full mesh of such tunnels, or more likely, a selection of tunnels connecting key SFFs to enable the construction of SFPs andtobalance load and traffic in the network. Further, the figure does not show anycontrollers:controllers; these would each have BGP connectivity to theClassifierclassifier and all of the SFFs.</t> <figureanchor="SFCarch" title="The SFCanchor="SFCarch"> <name>The Service Function Chaining Architecture ReferenceModel"> <artwork> <![CDATA[Model</name> <artwork name="" type="" align="left" alt=""><![CDATA[ Packets | | | ------------ | | | Classifier | | | ------+----- | ---+--- --------- ------- | | Tunnel | | | | | SFF-1 |===============| SFF-2 |=========| SFF-4 | | | | | | | | | -+-----+- | | | | ,,,,,,,,,,,,,,/,, \ | | | | ' .........../. ' ..\...... | | | | ' : SFb / : ' : \ SFc : | | | | ' : ---+- : ' : --+-- : | | | | ' : -| SFI | : ' : | SFI | : | | | | ' : -| ----- : ' : ----- : | | | | ' : | ----- : ' ......... | | | | ' : ----- : ' | | | | ' ............. ' | |--- Dests | | ' ' | |--- Dests | | ' ......... ' | | | | ' : ----- : ' | | | | ' : | SFI | : ' | | | | ' : --+-- : ' | | | | ' :SFd | : ' | | | | ' ....|.... ' | | | | ' | ' | | | | ' SFTx | ' | | | | ',,,,,,,,|,,,,,,,,' | | | | | | | | | ---+--- | | | | | | | | | |======| SFF-3 |====================| | ---+--- | | ---+--- | ------- | ....|.... ....|.... : | SFa: : | SFe: : --+-- : : --+-- : : | SFI | : : | SFI | : : ----- : : ----- : ......... .........]]> </artwork>]]></artwork> </figure> <t>As previously noted, <xref target="RFC8300"/>format="default"/> makes it clear that the mechanisms it defines are intended for use within a singleprovider'sprovider's operational domain. This reduces the requirements on the control plane function.</t> <t><xref target="RFC7665"/>section="5.2" sectionFormat="of"/> sets out the functions provided by a control plane foran SFC network in Section 5.2.a service function chaining network. The functions are broken down into sixitemsitems, the first four of which are completely covered by the mechanisms described in this document:<list style="numbers"> <t>Visibility</t> <ol spacing="normal" type="1"><li>Visibility of all SFs and the SFFs through which they arereached.</t> <t>Computationreached.</li> <li>Computation of SFPs and programming into thenetwork.</t> <t>Selectionnetwork.</li> <li>Selection of SFIs explicitly in the SFP or dynamically within thenetwork.</t> <t>Programmingnetwork.</li> <li>Programming of SFFs with forwarding pathinformation.</t> </list></t>information.</li> </ol> <t>The fifth andsixsixth items in the list in RFC 7665 concern the use of metadata. These are more peripheral to the control plane mechanisms defined in thisdocument,document but are discussed in <xref target="classy"/>.</t>format="default"/>.</t> </section> </section> <section anchor="sfcBGPRoutes"title="BGPnumbered="true" toc="default"> <name>BGP SFCRoutes">Routes</name> <t>This document defines a new AFI/SAFI for BGP, known as "SFC", with an NLRI that is described in this section.</t> <t>The format of the SFC NLRI is shown in <xref target="SFCnlriFig"/>.</t>format="default"/>.</t> <figureanchor="SFCnlriFig" title="Theanchor="SFCnlriFig"> <name>The Format of the SFCNLRI"> <artwork> <![CDATA[NLRI</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +---------------------------------------+ | Route Type (2 octets) | +---------------------------------------+ | Length (2 octets) | +---------------------------------------+ | Route Type specific (variable) | +---------------------------------------+]]> </artwork>]]></artwork> </figure> <t>TheRoute Type"Route Type" field determines the encoding of the rest of theroute typeRoute Type specific SFC NLRI.</t> <t>TheLength"Length" field indicates thelengthlength, inoctetsoctets, of theroute type specific"Route Type specific" field of the SFC NLRI.</t> <t>This document defines the following Route Types:<list style="numbers"> <t>Service</t> <ol spacing="normal" type="1"><li>Service Function Instance Route(SFIR)</t> <t>Service(SFIR)</li> <li>Service Function Path Route(SFPR)</t> </list></t> <t>A Service Function Instance Route (SFIR)(SFPR)</li> </ol> <t>An SFIR is used to identify an SFI.A Service Function Path Route (SFPR)An SFPR defines a sequence ofService FunctionsSFs (each of which has at least one instance advertised in an SFIR) that form an SFP.</t> <t>The detailed encoding and procedures for these Route Types are described in subsequent sections.</t> <t>The SFC NLRI is carried in BGP <xref target="RFC4271"/>format="default"/> using BGP Multiprotocol Extensions <xref target="RFC4760"/>format="default"/> with an Address Family Identifier (AFI) ofTBD131 and a Subsequent Address Family Identifier (SAFI) ofTBD2.9. TheNLRI"NLRI" field in the MP_REACH_NLRI/MP_UNREACH_NLRI attribute contains the SFC NLRI, encoded as specified above.</t> <t>In order for two BGP speakers to exchange SFC NLRIs, theyMUST<bcp14>MUST</bcp14> use BGPCapabilities Advertisementscapabilities advertisements to ensure that they both are capable of properly processing such NLRIs. This is done as specified in <xref target="RFC4760"/>,format="default"/>, by using capability code 1 (Multiprotocol BGP) with an AFI ofTBD131 and a SAFI ofTBD2.</t>9.</t> <t>Thenexthop"nexthop" field of the MP_REACH_NLRI attribute of the SFC NLRIMUST<bcp14>MUST</bcp14> be set to a loopback address of the advertising SFF.</t> <section anchor="sfiRoutes"title="Servicenumbered="true" toc="default"> <name>Service Function Instance Route(SFIR)">(SFIR)</name> <t><xref target="sfiRouteFig"/>format="default"/> shows the Route Type specific NLRI of the SFIR.</t> <figureanchor="sfiRouteFig" title="SFIRanchor="sfiRouteFig"> <name>SFIR Route Typespecific NLRI"> <artwork> <![CDATA[Specific NLRI</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Route Distinguisher (RD) (8 octets) | +--------------------------------------------+ | Service Function Type (2 octets) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t><xref target="RFC4364"/>format="default"/> defines a Route Distinguisher (RD)to consistas consisting of atwo byte Typetwo-byte "Type" field and asix byte Value fieldsix-byte "Value" field, and it defines RD types 0, 1, and 2. In this specification, the RD (used for the SFIR)MUST<bcp14>MUST</bcp14> be of type 0, 1, or 2.</t> <t>If two SFIRs are originated from different administrative domains (within the sameprovier'sprovider's operational domain), theyMUST<bcp14>MUST</bcp14> have different RDs. In particular, SFIRs from different VPNs (for different service function overlay networks)MUST<bcp14>MUST</bcp14> have different RDs, and those RDsMUST<bcp14>MUST</bcp14> be different from any non-VPN SFIRs.</t> <t>TheService Function TypeSFT identifies the functions/featuresa service functionan SF can offer, e.g.,Classifier,classifier, firewall, load balancer. There may be several SFIs that can perform a givenService Function.service function. Each node hosting an SFIMUST<bcp14>MUST</bcp14> originate an SFIR for each type of SF that it hosts (as indicated by the SFT value), and itMAY<bcp14>MAY</bcp14> advertise an SFIR for each instance of each type of SF.TheA minimal advertisement allows construction of valid SFPs and leaves the selection of SFIs to the local SFF;thea detailed advertisement may have scalingconcerns,concerns but allows aControllercontroller that constructs an SFP to make an explicit choice of SFI.</t> <t>Note that a node may advertise all its SFIs of one SFT in one shot using normal BGPUpdateUPDATE packing. That is, all of the SFIRs in an Update share a common Tunnel Encapsulation and Route Target (RT) attribute. See also <xref target="sfpatt"/>.</t>format="default"/>.</t> <t>The SFIR representing a given SFI will contain an NLRI withRD"RD" field set to an RD as specified above, and withSFTthe "SFT" field set to identify thatSFI's Service Function Type.SFI's SFT. The values for theSFT"SFT" field are taken from a registry administered by IANA (see <xref target="iana"/>).format="default"/>). A BGPUpdateUPDATE containing one or more SFIRsMUST<bcp14>MUST</bcp14> also include aTunnel Encapsulationtunnel encapsulation attribute <xreftarget="I-D.ietf-idr-tunnel-encaps" />.target="RFC9012" format="default"/>. If a data packet needs to be sent to an SFI identified in one of the SFIRs, it will be encapsulated as specified by theTunnel Encapsulation attribute,tunnel encapsulation attribute and then transmitted through the underlay network.</t> <t>Note that theTunnel Encapsulationtunnel encapsulation attributeMUST<bcp14>MUST</bcp14> contain sufficient information to allow the advertising SFF to identify the overlay or VPN networkwhichthat a received packet is transiting. This is because the [SPI, SI] in a received packet is specific to a particular overlay or VPN network.</t> <section anchor="poolid"title="SFIRnumbered="true" toc="default"> <name>SFIR Pool Identifier ExtendedCommunity">Community</name> <t>This document defines a new transitiveextended communityExtended Community <xref target="RFC4360"/>format="default"/> of typeTBD60x0b called theSFC extended community."SFC Extended Community". When used with Sub-Type 1, this is called theSFIR"SFIR Pool Identifier extendedcommunity.community". ItMAY<bcp14>MAY</bcp14> be included in SFIR advertisements, and it is used to indicate the identity of a pool of SFIRs to which an SFIR belongs. Since an SFIR may be a member ofmultiple pools,more than one pool, multiple of these extended communities may be present on a single SFIR advertisement.</t> <t>SFIR pools allow SFIRs to be grouped for any purpose. Possible uses include control plane scalability and stability. A pool identifier may be included in an SFPR to indicate a set of SFIs that are acceptable at a specific point on an SFP (see Sections <xref format="counter" target="sfttlv"/> and <xreftarget="SFPR"/>).</t>target="SFPR" format="counter"/>).</t> <t>The SFIR Pool Identifierextended communityExtended Community is encoded in 8 octets as shown in <xreftarget="poolFig"/>.</t>target="poolFig" format="default"/>.</t> <figureanchor="poolFig" title="Theanchor="poolFig"> <name>The SFIR Pool Identifier ExtendedCommunity"> <artwork> <![CDATA[Community</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Type =TBD60x0b (1 octet) | +--------------------------------------------+ | Sub-Type = 1 (1 octet) | +--------------------------------------------+ | SFIR Pool IdentifierValuevalue (6 octets) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t>The SFIR Pool IdentifierValuevalue is encoded in a6 octet6-octet field in network byte order, and the value is unique within the scope of an overlay network. This means that pool identifiers need to be centrally managed, which is consistent with the assignment of SFIs to pools.</t> </section> <section anchor="swapnstack"title="MPLSnumbered="true" toc="default"> <name>MPLS Mixed Swapping/Stacking ExtendedCommunity">Community</name> <t>As noted in <xref target="poolid"/>,format="default"/>, this document defines a new transitiveextended communityExtended Community of typeTBD60x0b called theSFC extended community."SFC Extended Community". When used with Sub-Type 2, this is called theMPLS"MPLS Mixed Swapping/Stacking Labelsextended community.Extended Community". The community is encoded as shown in <xreftarget="swapFig"/>.target="swapFig" format="default"/>. It contains a pair of MPLS labels: an SFC Context Label and an SFLabelLabel, as described in <xreftarget="RFC8595"/>.target="RFC8595" format="default"/>. Each label is 20 bits encoded in a 3-octet(24 bit)(24-bit) field with 4 trailing bits thatMUST<bcp14>MUST</bcp14> be set to zero.</t> <figureanchor="swapFig" title="Theanchor="swapFig"> <name>The MPLS Mixed Swapping/Stacking Labels ExtendedCommunity"> <artwork> <![CDATA[Community</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Type =TBD60x0b (1 octet) | +--------------------------------------------| | Sub-Type = 2 (1 octet) | +--------------------------------------------| | SFC Context Label (3 octets) | +--------------------------------------------| | SF Label (3 octets) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t>Note that it is assumed that each SFF has one or more globally unique SFC Context Labels and that thecontext labelcontext-label space and theSPI addressSPI-address space aredisjoint (i.e.,disjoint. In other words, a label value cannot be usedbothto indicate both an SFC context and an SPI, and it can be determined from knowledge of the label spaces whether a label indicates an SFC context or anSPI).</t>SPI.</t> <t>If an SFF supports SFP Traversal with an MPLS LabelStackStack, itMUST<bcp14>MUST</bcp14> include thisextended communityExtended Community with the SFIRs that it advertises.</t> <t>See <xreftarget="swapOp"/>target="swapOp" format="default"/> for a description of how thisextended communityExtended Community is used.</t> </section> </section> <section anchor="sfpRoutes"title="Servicenumbered="true" toc="default"> <name>Service Function Path Route(SFPR)">(SFPR)</name> <t><xref target="sfpRouteFig"/>format="default"/> shows the Route Type specific NLRI of the SFPR.</t> <figureanchor="sfpRouteFig" title="SFPRanchor="sfpRouteFig"> <name>SFPR Route Type SpecificNLRI"> <artwork> <![CDATA[NLRI</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +-----------------------------------------------+ | Route Distinguisher (RD) (8 octets) | +-----------------------------------------------+ | Service Path Identifier (SPI) (3 octets) | +-----------------------------------------------+]]> </artwork>]]></artwork> </figure> <t><xref target="RFC4364"/>format="default"/> defines a Route Distinguisher (RD)to consistas consisting of atwo byte Typetwo-byte "Type" field and asix byte Value fieldsix-byte "Value" field, and it defines RD types 0, 1, and 2. In this specification, the RD (used for the SFPR)MUST<bcp14>MUST</bcp14> be of type 0, 1, or 2.</t> <t>All SFPsMUST<bcp14>MUST</bcp14> be associated with an RD. The association of an SFP with an RD is determined by provisioning. If two SFPRs are originated from differentControllerscontrollers, theyMUST<bcp14>MUST</bcp14> have different RDs. Additionally, SFPRs from different VPNs (i.e., in different service function overlay networks)MUST<bcp14>MUST</bcp14> have different RDs, and those RDsMUST<bcp14>MUST</bcp14> be different from any non-VPN SFPRs.</t> <t>The ServicePath Identifierpath identifier is defined in <xref target="RFC8300"/>format="default"/> and is the value to be placed in theService"Service PathIdentifierIdentifier" field of the NSHheaderof any packet sent on thisService Function Path.SFP. It is expected that one or moreControllerscontrollers will originate these routes in order to configure a service function overlay network.</t> <t>The SFP is described in a new BGP Path attribute, the SFP attribute. <xref target="sfpatt"/>format="default"/> shows the format of that attribute.</t> <section anchor="sfpatt"title="Thenumbered="true" toc="default"> <name>The SFPAttribute">Attribute</name> <t><xref target="RFC4271"/>format="default"/> defines BGP Path attributes. This document introduces a new Optional Transitive Path attribute called theSFP attribute"SFP attribute", with valueTBD3 to be assigned by IANA.37. The first SFP attributeMUST<bcp14>MUST</bcp14> beprocessedprocessed, and subsequent instancesMUST<bcp14>MUST</bcp14> be ignored.</t> <t>The common fields of the SFP attribute are set as follows:<list style="symbols"> <t>Optional</t> <ul spacing="normal"> <li>The Optional bit is set to 1 to indicate that this is an optionalattribute.</t> <t>Theattribute.</li> <li>The Transitive bit is set to 1 to indicate that this is a transitiveattribute.</t> <t>Theattribute.</li> <li>The Extended Length bit is set if the length of the SFP attribute is encoded in one octet (set to 0) or two octets (set to1)1), as described in <xref target="RFC4271"/>.</t> <t>Theformat="default"/>.</li> <li>The Attribute Type Code is set toTBD3.</t> </list></t>37.</li> </ul> <t>The content of the SFP attribute is a series of Type-Length-Value (TLV) constructs. Some TLVs may includesub-TLVs.Sub-TLVs. All TLVs andsub-TLVsSub-TLVs have a commonformat that is: <list style="symbols"> <t>Type:format: </t> <dl> <dt>Type:</dt><dd> A single octet indicating the type of the SFP attribute TLV. Values are taken from the registry described in <xref target="ianasftlv"/>.</t> <t>Length:format="default"/>.</dd> <dt>Length:</dt><dd> Atwo octettwo-octet field indicating the length of the data following theLength field"Length" field, counted inoctets.</t> <t>Value:octets.</dd> <dt>Value:</dt><dd> The contents of theTLV.</t> </list></t>TLV.</dd> </dl> <t>The formats of the TLVs defined in this document are shown in the following sections. The presence rules and meanings are as follows.<list style="symbols"> <t>The</t> <ul spacing="normal"> <li>The SFP attribute contains a sequence of zero or more Association TLVs. That is, the Association TLV isOPTIONAL.<bcp14>OPTIONAL</bcp14>. Each Association TLV provides an association between this SFPR and another SFPR. Each associated SFPR is indicated using the RD with which it is advertised (we say the SFPR-RD to avoidambiguity).</t> <t>Theambiguity).</li> <li>The SFP attribute contains a sequence of one or more Hop TLVs. Each Hop TLV contains all of the information about a single hop in theSFP.</t> <t>EachSFP.</li> <li>Each Hop TLV contains an SI value and a sequence of one or more SFT TLVs. Each SFT TLV contains an SFI reference for each instance of an SF that is allowed at this hop of the SFP for the specific SFT. Each SFI is indicated using the RD with which it is advertised (we say the SFIR-RD to avoidambiguity).</t> </list></t> <t>Section 6 of <xref target="RFC4271"/>ambiguity).</li> </ul> <t><xref target="RFC4271" section="6" sectionFormat="of"/> describes the handling of malformed BGP attributes, or those that are in error in some way. <xref target="RFC7606"/>format="default"/> revises BGP error handling specifically for the UPDATE message, provides guidelines for the authors of documents defining new attributes, and revises theerror handlingerror-handling procedures for a number of existing attributes. This document introduces the SFP attribute and so defines error handling as follows:<list style="symbols"> <t>When</t> <ul spacing="normal"> <li>When parsing a message, an unknown Attribute TypecodeCode or a length that suggests that the attribute is longer than the remaining message is treated as a malformedmessagemessage, and the "treat-as-withdraw" approach is used as per <xreftarget="RFC7606"/>.</t>target="RFC7606" format="default"/>.</li> <li> <t>When parsing a message that contains an SFP attribute, the following cases constitute errors:<list style="numbers"> <t>Optional</t> <ol spacing="normal" type="1"><li>Optional bit is set to 0 in the SFPattribute.</t> <t>Transitiveattribute.</li> <li>Transitive bit is set to 0 in the SFPattribute.</t> <t>Unknown TLV typeattribute.</li> <li>Unknown "TLV Type" field found in the SFPattribute.</t> <t>TLVattribute.</li> <li>TLV length that suggests the TLV extends beyond the end of the SFPattribute.</t> <t>Associationattribute.</li> <li>Association TLV contains an unknownSFPR-RD.</t> <t>NoSFPR-RD.</li> <li>No Hop TLV found in the SFPattribute.</t> <t>No sub-TLVattribute.</li> <li>No Sub-TLV found in a HopTLV.</t> <t>UnknownTLV.</li> <li>Unknown SFIR-RD found in an SFTTLV.</t> </list></t>TLV.</li> </ol> </li> <li> <t>The errors listed above are treated as follows:<list style="hanging"> <t hangText="1., 2., 4., 6., 7.:">The</t> <dl newline="false" spacing="normal"> <dt>1, 2, 4, 6, 7:</dt> <dd>The attributeMUST<bcp14>MUST</bcp14> be treated as malformed and the "treat-as-withdraw" approach used as per <xreftarget="RFC7606"/>.</t> <t hangText="3.:">Unknowntarget="RFC7606" format="default"/>.</dd> <dt>3:</dt> <dd>Unknown TLVsMUST<bcp14>MUST</bcp14> be ignored, and message processingMUST continue.</t> <t hangText="5., 8.:">The<bcp14>MUST</bcp14> continue.</dd> <dt>5, 8:</dt> <dd>The absence of an RD with which to correlate is nothing more than a soft error. The receiverSHOULD<bcp14>SHOULD</bcp14> store the information from the SFP attribute until a corresponding advertisement isreceived.</t> </list></t> </list></t>received.</dd> </dl> </li> </ul> <section anchor="assoctlv"title="Thenumbered="true" toc="default"> <name>The AssociationTLV">TLV</name> <t>The Association TLV is an optional TLV in the SFP attribute. ItMAY<bcp14>MAY</bcp14> be present multiple times. Each occurrence provides an association with another SFP as advertised in another SFPR. The format of the Association TLV is shown in <xref target="assoctlvfig"/></t>format="default"/>.</t> <figureanchor="assoctlvfig" title="Theanchor="assoctlvfig"> <name>The Format of the AssociationTLV"> <artwork> <![CDATA[TLV</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Type = 1 (1 octet) | +--------------------------------------------| | Length (2 octets) | +--------------------------------------------| | Association Type (1 octet) | +--------------------------------------------| | Associated SFPR-RD (8 octets) | +--------------------------------------------| | Associated SPI (3 octets) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t>The fields are as follows:<list style="emtpy" > <t>Type</t> <ul spacing="normal"> <li>"Type" is set to 1 to indicate an AssociationTLV.</t> <t>LengthTLV.</li> <li>"Length" indicates the length in octets of theAssociation Type"Association Type" andAssociated SFPR-RD"Associated SFPR-RD" fields. The value of theLength"Length" field is12.</t> <t>The Association Type12.</li> <li>The "Association Type" fieldindicateindicates the type of association. The values are tracked in an IANA registry (see <xref target="ianaassoc"/>).format="default"/>). Only one value is defined in this document:typeType 1 indicates association of two unidirectional SFPs to form a bidirectional SFP. An SFP attributeSHOULD NOT<bcp14>SHOULD NOT</bcp14> contain more than one Association TLV with Association Type1:1; if more than one is present, the first oneMUST<bcp14>MUST</bcp14> beprocessedprocessed, and subsequent instancesMUST<bcp14>MUST</bcp14> be ignored. Note that documents that define newAssociation Typesassociation types must also define the presence rules for Association TLVs of the newtype.</t> <t>Thetype.</li> <li>The Associated SFPR-RD contains the RD of the associated SFP as advertised in anSFPR.</t> <t>TheSFPR.</li> <li>The Associated SPI contains the SPI of the associated SFP as advertised in anSFPR.</t> </list></t>SFPR.</li> </ul> <t>Association TLVs with unknown Association Type valuesSHOULD<bcp14>SHOULD</bcp14> be ignored. Association TLVs that contain an Associated SFPR-RD value equal to the RD of the SFPR in which they are containedSHOULD<bcp14>SHOULD</bcp14> be ignored. If the Associated SPI is not equal to the SPI advertised in the SFPR indicated by the AssociatedSFPR-RDSFPR-RD, then the Association TLVSHOULD<bcp14>SHOULD</bcp14> be ignored. In all three of thesecasescases, an implementationMAY<bcp14>MAY</bcp14> reject the SFP attribute as malformed and use the "treat-as-withdraw" approach per <xreftarget="RFC7606"/>, however implementerstarget="RFC7606" format="default"/>; however, implementors are cautioned that such an approach may make an implementation less flexible in the event of future extensions to this protocol.</t> <t>Note that when two SFPRs reference each other using the Association TLV, one SFPR advertisement will be received before the other. Therefore, processing of an associationMUST NOT<bcp14>MUST NOT</bcp14> be rejected simply because the Associated SFPR-RD is unknown.</t> <t>Further discussion of correlation of SFPRs is provided in <xref target="correlation"/>.</t>format="default"/>.</t> </section> <section anchor="hoptlv"title="Thenumbered="true" toc="default"> <name>The HopTLV">TLV</name> <t>There is one Hop TLV in the SFP attribute for each hop in the SFP. The format of the Hop TLV is shown in <xref target="hoptlvfig"/>.format="default"/>. At least one Hop TLVMUST<bcp14>MUST</bcp14> be present in an SFP attribute.</t> <figureanchor="hoptlvfig" title="Theanchor="hoptlvfig"> <name>The Format of the HopTLV"> <artwork> <![CDATA[TLV</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Type = 2 (1 octet) | +--------------------------------------------| | Length (2 octets) | +--------------------------------------------| | Service Index (1 octet) | +--------------------------------------------| | Hop Details (variable) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t>The fields are as follows:<list style="emtpy" > <t>Type</t> <ul spacing="normal"> <li>"Type" is set to 2 to indicate a HopTLV.</t> <t>LengthTLV.</li> <li>"Length" indicates thelengthlength, inoctetsoctets, of theService Index"Service Index" andHop Details fields.</t> <t>The"Hop Details" fields.</li> <li>The Service Index is defined in <xref target="RFC8300"/>format="default"/> and is the value found in theService Index"Service Index" field of the NSHheaderthat an SFF will use tolookuplook up to which next SFI a packet is to besent.</t> <t>The Hop Detailssent.</li> <li>The "Hop Details" field consists of a sequence of one or moresub-TLVs.</t> </list></t>Sub-TLVs.</li> </ul> <t>Each hop of the SFP may demand that a specific type of SF is executed, and that type is indicated insub-TLVsSub-TLVs of the Hop TLV. At least onesub-TLV MUSTSub-TLV <bcp14>MUST</bcp14> be present. This document defines the SFT Sub-TLV (see <xref target="sfttlv"/>format="default"/>) and the MPLS Swapping/Stacking Sub-TLV (seeSection<xref target="swapTLV"/>:format="default"/>); othersub-TLVsSub-TLVs may be defined in future.ThisThe SFT Sub-TLV provides a list of which types of SF are acceptable at a specific hop, and for each type it allows a degree of control to be imposed on the choice of SFIs of that particulartype.</t>type. The MPLS Swapping/Stacking Sub-TLV indicates the type of SFC encoding to use in an MPLS label stack. </t> <t>If no Hop TLV is present in an SFPAttribute,attribute, it is a malformedattribute</t>attribute.</t> </section> <section anchor="sfttlv"title="Thenumbered="true" toc="default"> <name>The SFTSub-TLV">Sub-TLV</name> <t>The SFT Sub-TLVMAY<bcp14>MAY</bcp14> be included in the list ofsub-TLVsSub-TLVs of the Hop TLV. The format of the SFT Sub-TLV is shown in <xref target="sfttlvfig"/>.format="default"/>. The Hop Sub-TLV contains a list of SFIR-RD values each taken from the advertisement of an SFI. Together they form a list of acceptable SFIs of the indicated type.</t> <figureanchor="sfttlvfig" title="Theanchor="sfttlvfig"> <name>The Format of the SFTSub-TLV"> <artwork> <![CDATA[Sub-TLV</name> <artwork name="" type="" align="left" alt=""><![CDATA[ +--------------------------------------------+ | Type = 3 (1 octet) | +--------------------------------------------| | Length (2 octets) | +--------------------------------------------| | Service Function Type (2 octets) | +--------------------------------------------| | SFIR-RD List (variable) | +--------------------------------------------+]]> </artwork>]]></artwork> </figure> <t>The fields are as follows:<list style="emtpy" > <t>Type</t> <ul> <li>"Type" is set to 3 to indicate an SFTSub-TLV.</t> <t>LengthSub-TLV.</li> <li>"Length" indicates thelengthlength, inoctetsoctets, of theService"Service FunctionTypeType" andSFIR-RD List fields.</t> <t>The Service Function Type"SFIR-RD List" fields.</li> <li>The SFT value indicates the category (type) of SF that is to be executed at this hop. The types are as advertised for the SFs supported by the SFFs. SFT values in the range 1-31 areSpecial Purposespecial-purpose SFT values and have meanings defined by the documents that describe them--- the value'Change Sequence'"Change Sequence" is defined in <xref target="changeseq"/>format="default"/> of thisdocument.</t> <t>Thedocument.</li> <li>The hop description is further qualified beyond the specification of the SFTs by listing, for each SFT in each hop, the SFIs that may be used at the hop. The SFIs are identified using the SFIR-RDs from the advertisements of the SFIs in the SFIRs. Note that if the list contains one or more SFIR Pool Identifiers, then foreacheach, the SFIR-RD list is effectively expanded to include the SFIR-RD of each SFIR advertised with that SFIR Pool Identifier. An SFIR-RD of value zero has specialmeaningmeaning, as described in <xref target="selection"/>.format="default"/>. Each entry in the list is eight octets long, and the number of entries in the list can be deduced from the value of theLength field.</t> <t>Note"Length" field.</li> <li>Note that an SFIR-RD is of type 0, 1, or 2 (as described in <xref target="sfiRoutes"/>. Thusformat="default"/>). Thus, thehigh orderhigh-order octet of an RD found in an SFIR-RD List always has a value of 0x00. However, thehigh orderhigh-order octet of an SFIR Pool Identifier (anextended communityExtended Community withType"Type" fieldTBD6),0x0b) will always have anon-zerononzero value. This allows the node processing the SFIR-RDListlist to distinguish between the two types of listentry.</t> </list></t>entry.</li> </ul> </section> <section anchor="swapTLV"title="MPLSnumbered="true" toc="default"> <name>MPLS Swapping/StackingSub-TLV">Sub-TLV</name> <t>The MPLS Swapping/Stacking Sub-TLV (Type value 4) is azero length sub-TLVzero-length Sub-TLV that isOPTIONAL<bcp14>OPTIONAL</bcp14> in the Hop TLV and is used when the data representation is MPLS (see <xreftarget="representation"/>).target="representation" format="default"/>). Whenpresentpresent, it indicates to theClassifierclassifier imposing an MPLS label stack that the current hop is to use an {SFC Context Label, SF label} rather than an {SPI, SF} label pair. See <xreftarget="swapOp"/>target="swapOp" format="default"/> for more details.</t> </section> <section anchor="sfpTraverse"title="SFPnumbered="true" toc="default"> <name>SFP Traversal With MPLS Label StackTLV">TLV</name> <t>The SFP Traversal With MPLS Label Stack TLV (Type value 5) is azero lengthzero-length TLV that can be carried in the SFPAttributeattribute and indicates to theClassifierclassifier and the SFFs on the SFP that an MPLS label stack with label swapping/stacking is to be used for packets traversing the SFP. All of the SFFs specified at each of theSFP'sSFP's hopsMUST<bcp14>MUST</bcp14> have advertised an MPLS Mixed Swapping/Stacking Extended Community (see <xref target="swapnstack"/>)format="default"/>) for the SFP to be considered usable.</t> </section> </section> <section anchor="sfparules"title="Generalnumbered="true" toc="default"> <name>General RulesFor Thefor the SFPAttribute">Attribute</name> <t>It is possible for the same SFI, as described by an SFIR, to be used in multiple SFPRs.</t> <t>When two SFPRs have the same SPI but differentSFPR-RDsSFPR-RDs, there can be three cases:<list style="symbols"> <t>Two</t> <ol type="1" spacing="normal"> <li>Two or moreControllerscontrollers are originating SFPRs for the same SFP. In thiscasecase, the content of the SFPRs isidenticalidentical, and the duplication is to ensure receipt andtoprovideController redundancy.</t> <t>Therecontroller redundancy.</li> <li>There is a transition in content of the advertisedSFPSFP, and the advertisements may originate from one or moreControllers.controllers. In thiscasecase, the content of the SFPRs will bedifferent.</t> <t>Thedifferent.</li> <li>The reuse of an SPI may result from a configurationerror.</t> </list></t> <t>In all cases, thereerror.</li> </ol> <t>There is no way in any of these cases for the receiving SFF to know which SFPR to process, and the SFPRs could be received in any order. At any point in time, when multiple SFPRs have the same SPI but different SFPR-RDs, the SFFMUST<bcp14>MUST</bcp14> use the SFPR with the numerically lowest SFPR-RD when interpreting the RDs as 8-octet integers in network byte order. The SFFSHOULD<bcp14>SHOULD</bcp14> log this occurrence to assist with debugging.</t> <t>Furthermore, aControllercontroller that wants to change the content of an SFP isRECOMMENDED<bcp14>RECOMMENDED</bcp14> to use a new SPI and so create a new SFP onto which theClassifiersclassifiers can transition packet flows before the SFPR for the old SFP is withdrawn. This avoids any race conditions with SFPR advertisements.</t> <t>Additionally, aController SHOULD NOT re-usecontroller <bcp14>SHOULD NOT</bcp14> reuse an SPI after it has withdrawn the SFPR that used it until at least a configurable amount of time has passed. This timerSHOULD<bcp14>SHOULD</bcp14> have a default of one hour.</t> </section> </section> </section> <section anchor="mode"title="Modenumbered="true" toc="default"> <name>Mode ofOperation">Operation</name> <t>This document describes the use of BGP as a control plane to create and manage a service function overlay network.</t> <section anchor="rt"title="Route Targets">numbered="true" toc="default"> <name>Route Targets</name> <t>The main feature introduced by this document is the ability to create multiple service function overlay networks through the use of Route Targets (RTs) <xref target="RFC4364"/>.</t>format="default"/>.</t> <t>Every BGP UPDATE containing an SFIR or SFPR carries one or more RTs. The RT carried by a particular SFIR or SFPR is determined by the provisioning of theroute'sroute's originator.</t> <t>Every node in a service function overlay network is configured with one or more import RTs. Thus, each SFF will import only the SFPRs with matchingRTsRTs, allowing the construction of multiple service function overlay networks or the instantiation ofService Function ChainsSFCs withinana Layer 3 Virtual Private Network (L3VPN) or Ethernet VPN (EVPN) instance (see <xref target="private"/>).format="default"/>). An SFF that has a presence in multiple service function overlay networks (i.e., one that imports more than one RT) will usually maintain separate forwarding state for each overlay network.</t> </section> <section anchor="SFIR"title="Servicenumbered="true" toc="default"> <name>Service Function InstanceRoutes">Routes</name> <t>The SFIR (see <xref target="sfiRoutes"/>)format="default"/>) is used to advertise the existence and location of a specificService Function Instance andSFI; it consists of:<list style="symbols"> <t>The</t> <ul spacing="normal"> <li>The RT as justdescribed.</t> <t>Adescribed.</li> <li>A Service Function Type (SFT) that is the type of service function that is provided (such as"firewall").</t> <t>A"firewall").</li> <li>A Route Distinguisher (RD) that is unique to a specificoverlay.</t> </list></t>overlay.</li> </ul> </section> <section anchor="SFPR"title="Servicenumbered="true" toc="default"> <name>Service Function PathRoutes">Routes</name> <t>The SFPR (see <xref target="sfpRoutes"/>)format="default"/>) describes a specific path ofa Service Function Chain. Thean SFC. The SFPR contains the Service Path Identifier (SPI) used to identify the SFP in the NSH in the data plane. It also contains a sequence of Service Indexes (SIs). Each SI identifies a hop in the SFP, and each hop is a choice between oneofor more SFIs.</t> <t>As described in this document, eachService Function Path RouteSFP route is identified in the service function overlay network by an RD and an SPI. The SPI is unique within a single VPN instance supported by the underlay network.</t> <t>The SFPR advertisement comprises:<list style="symbols"> <t>An</t> <ul spacing="normal"> <li>An RT as described in <xref target="rt"/>.</t>format="default"/>.</li> <li> <t>A tuple that identifies theSFPR <list style="symbols"> <t>AnSFPR. </t> <ul spacing="normal"> <li>An RD that identifies an advertisement of anSFPR.</t> <t>TheSFPR.</li> <li>The SPI that uniquely identifies this path within the VPN instance distinguished by the RD. This SPI also appears in theNSH.</t> </list></t> <t>ANSH.</li> </ul> </li> <li>A series ofService Indexes.SIs. Each SI is used in the context of a particular SPI and identifies one or more SFs (distinguished by theirSFTs) and forSFTs). For eachSFSF, it identifies a set of SFIs that instantiate the SF. The values of the SI indicate the order in which the SFs are to be executed in the SFP that is represented by theSPI.</t> <t>TheSPI.</li> <li>The SI is used in the NSH to identify the entries in the SFP. Note that the SI values have meaning only relative to a specific path. They have no semantic other than to indicate the order ofService FunctionsSFs within the path and are assumed to be monotonically decreasing from the start to the end of the path <xref target="RFC8300"/>.</t>format="default"/>.</li> <li> <t>EachService IndexSI is associated with a set of one or moreService Function InstancesSFIs that can be used to provide the indexedService FunctionSF within the path. Each member of the set comprises:<list style="symbols"> <t>The</t> <ul spacing="normal"> <li>The RD used in an SFIR advertisement of theSFI.</t> <t>TheSFI.</li> <li>The SFT that indicates the type of function as used in the same SFIR advertisement of theSFI.</t> </list></t> </list></t>SFI.</li> </ul> </li> </ul> <t>This may be summarized asfollowsfollows, where the notations "SFPR-RD" and "SFIR-RD" are used to distinguish the two different RDs, and where "*" indicates a multiplier:<list style="empty"> <t>RT,</t> <artwork> RT, {SFPR-RD, SPI}, m * {SI, {n * {SFT, p * SFIR-RD} }}</t> </list></t>} </artwork> <t>Where:<list style="empty"> <t>RT:</t> <dl> <dt>RT:</dt><dd> RouteTarget</t> <t>SFPR-RD: TheTarget</dd> <dt>SFPR-RD:</dt><dd>The Route Descriptor of theService Function Path Route advertisement</t> <t>SPI: ServiceSFPR advertisement</dd> <dt>SPI:</dt><dd>Service Path Identifier used in theNSH</t> <t>m: TheNSH</dd> <dt>m:</dt><dd>The number of hops in theService Function Path</t> <t>n: TheSFP</dd> <dt>n:</dt><dd>The number of choices ofService Function TypeSFT for a specifichop</t> <t>p: Thehop</dd> <dt>p:</dt><dd>The number of choices ofService Function InstanceSFI for a givenService Function TypeSFT in a specifichop</t> <t>SI: Servicehop</dd> <dt>SI:</dt><dd>Service Index used in the NSH to indicate a specifichop</t> <t>SFT: Thehop</dd> <dt>SFT:</dt><dd>The Service Function Type used in the same advertisement of theService Function Instance Route</t> <t>SFIR-RD: TheSFIR </dd> <dt>SFIR-RD:</dt><dd>The Route Descriptor used in an advertisement of theService Function Instance Route</t> </list></t>SFIR</dd> </dl> <t>That is, there can be multiple SFTs at a givenhophop, as described in <xreftarget="selection"/>.</t>target="selection" format="default"/>.</t> <t>Note that the values of SI are from the set {255, ..., 1} and are monotonically decreasing within the SFP. SIsMUST<bcp14>MUST</bcp14> appear in order within the SFPR (i.e., monotonically decreasing) andMUST NOT<bcp14>MUST NOT</bcp14> appear more than once. GapsMAY<bcp14>MAY</bcp14> appear in thesequencesequence, as described in <xref target="lacunae"/>.format="default"/>. Malformed SFPRsMUST<bcp14>MUST</bcp14> be discarded andMUST<bcp14>MUST</bcp14> cause any previous instance of the SFPR (same SFPR-RD and SPI) to be discarded.</t> <t>Note that if the SFIR-RD list in an SFT TLV contains one or more SFIR Poolidentifiers,Identifiers, then in the above expression,'p'"p" is the sum of the number of individual SFIR-RD values and the sum for each SFIR Pool Identifier of the number of SFIRs advertised with that SFIR Pool Identifier.I.e.,In other words, the list of SFIR-RD values is effectively expanded to include the SFIR-RD of each SFIR advertised with each SFIR Pool Identifier in the SFIR-RD list.</t> <t>The choice of SFI is explained further in <xref target="selection"/>.format="default"/>. Note that an SFIR-RD value of zero has specialmeaningmeaning, as described in thatSection.</t>section.</t> </section> <section anchor="classy"title="Classifier Operation">numbered="true" toc="default"> <name>Classifier Operation</name> <t>As shown in <xref target="SFCarch"/>,format="default"/>, theClassifierclassifier is a component that is used to assign packets to an SFP.</t> <t>TheClassifierclassifier is responsible for determining to which packet flow a packet belongs. The mechanism it uses to achieve that classification is out of the scope of thisdocument,document but might include inspection of the packet header. TheClassifierclassifier has been instructed (by theControllercontroller or through some other configuration mechanism--- see <xref target="fspecclassy"/>)format="default"/>) which flows are to be assigned to which SFPs, and so it can impose an NSH on each packet and initialize the NSH with the SPI of the selected SFP and the SI of its first hop.</t> <t>Note that instructions delivered to theClassifierclassifier may include information about the metadata to encode (and the format for that encoding) on packets that are classified by theClassifierclassifier to a particular SFP. As mentioned in <xref target="ctrlover"/>,format="default"/>, this corresponds to the fifth element of control plane functionality described in <xref target="RFC7665"/>.format="default"/>. Such instructions fall outside the scope of this specification(although,(but see <xref target="fspecclassy"/>),format="default"/>), as do instructions to otherSFCservice function chaining elements on how to interpret metadata (as described in the sixth element of control plane functionality described in <xref target="RFC7665"/>.</t>format="default"/>).</t> </section> <section anchor="SFF"title="Servicenumbered="true" toc="default"> <name>Service Function ForwarderOperation">Operation</name> <t>Each packet sent to an SFF is transmitted encapsulated in an NSH. The NSH includes an SPI and SI: the SPI indicates the SFPR advertisement that announced theService Function Path;SFP; the tuple SPI/SI indicates a specific hop in a specific path and maps to the RD/SFT of a particular SFIR advertisement.</t> <t>When an SFF gets an SFPRadvertisementadvertisement, it will first determine whether to import the route by examining the RT. If the SFPR isimportedimported, the SFF then determines whether it is on the SFP by looking for its own SFIR-RDs or any SFIR-RD with value zero in the SFPR. For each occurrence in the SFP, the SFF creates forwarding state for incoming packets and forwarding state for outgoing packets that have been processed by the specified SFI.</t> <t>The SFF creates local forwarding state for packets that it receives from other SFFs. This state makes the association between the SPI/SI in the NSH of the received packet and one or more specific localSFIsSFIs, as identified by the SFIR-RD/SFT. If there are multiple local SFIs thatmatchmatch, this is because a single advertisement was made for a set of equivalentSFIsSFIs, and the SFF may use local policy (such as load balancing) to determine to which SFI to forward a received packet.</t> <t>The SFF also createsnext hopnext-hop forwarding state for packets received back from the local SFI that need to be forwarded to the next hop in the SFP. There may be a choice of nexthopshops, as described in <xref target="SFPR"/>.format="default"/>. The SFF could install forwarding state for all potential nexthops,hops or it could choose to only install forwarding statetofor a subset of the potential next hops. If a choice ismademade, then it will be as described in <xref target="selection"/>.</t>format="default"/>.</t> <t>The installed forwarding state may change overtimetime, reacting to changes in the underlay network and the availability of particular SFIs. Note that the forwarding state describes how one SFFsendsends packets to another SFF, but not how those packets are routed through the underlay network. SFFs may be connected by tunnels across the underlay, or packets may be sent addressed to the next SFF and routed through the underlay. In any case, transmission across the underlay requires encapsulation of packets with a header for transport in the underlay network.</t> <t>Note that SFFs only create and store forwarding state for the SFPs on which they are included. They do not retain state for all SFPs advertised.</t> <t>An SFF may also install forwarding state to support looping, jumping, and branching. The protocol mechanism for explicit control of looping, jumping, and branching uses a specific reserved SFT value at a given hop of an SFPR and is described in <xref target="changeseq"/>.</t>format="default"/>.</t> <section anchor="lacunae"title="Processing With 'Gaps'numbered="true" toc="default"> <name>Processing with "Gaps" in the SISequence">Sequence</name> <t>The behavior of anSFSF, as described in <xref target="RFC8300"/>format="default"/>, is to decrement the value of theSI"SI" field in the NSH by one before returning a packet to the local SFF for further processing. This means that there is a good reason to assume that the SFP is composed of a series ofSFsSFs, each indicated by an SI value one less than the previous.</t> <t>However, there is an advantage to havingnon-successivenonsuccessive SIs in an SPI. Consider the case where an SPI needs to be modified by the insertion or removal of an SF. In the lattercasecase, this would lead to a "gap" in the sequence of SIs, and in the former case, this could only be achieved if a gap already existed into which the new SF with its new SI value could be inserted. Otherwise, all "downstream" SFs would need to be renumbered.</t> <t>Now, of course, such renumbering could be performed, but it would lead to a significant disruption to the SFC as all the SFFs along the SFP were "reprogrammed". Thus, to achieve dynamic modification of an SFP (andeven,even in-servicemodification)modification), it is desirable to be able to make these modifications without changing the SIs of the elements that were present before the modification. This will produce much more consistent/predictable behavior during the convergenceperiodperiod, where otherwise the change would need to be fully propagated.</t> <t>Another approach says that any change to an SFP simply creates a new SFP that can be assigned a new SPI. All that would be needed would be to give a new instruction to theClassifierclassifier, and traffic would be switched to the new SFP that contains the new set of SFs. This approach ispractical,practical but neglects to consider that the SFP may be referenced by other SFPs (through "branch" instructions) and used by manyClassifiers.classifiers. In thosecasescases, the corresponding configuration resulting from a change in SPI may have wide ripples andgivecreate scope for errors that are hard to trace.</t> <t>Therefore, while this document requires that the SI values in an SFP aremonotonicmonotonically decreasing, it makes no assumption that the SI values are sequential. Configuration tools may apply that rule, but they are not required to. To support this, an SFFSHOULD<bcp14>SHOULD</bcp14> process as follows when it receives a packet:<list style="symbols"> <t>If</t> <ul spacing="normal"> <li>If the SI indicates a known entry in the SFP, the SFFMUST<bcp14>MUST</bcp14> process the packet as normal, looking up the SI and determining to which local SFI to deliver thepacket.</t> <t>Ifpacket.</li> <li>If the SI does not match an entry in the SFP, the SFFMUST<bcp14>MUST</bcp14> reduce the SI value to the next (smaller) value present in the SFP and process the packet using thatSI.</t> <t>IfSI.</li> <li>If there is no smaller SI (i.e., if the end of the SFP has beenreached)reached), the SFFMUST<bcp14>MUST</bcp14> treat the SI value asinvalidnot valid, as described in <xref target="RFC8300"/>.</t> </list>format="default"/>.</li> </ul> <t> This makes the behavior described in this document a superset of the function in <xref target="RFC8300"/>.format="default"/>. That is, an implementation that strictly follows RFC 8300 in performing SI decrements in units ofone,one is perfectly in line with the mechanisms defined in this document.</t> <t>SFF implementationsMAY<bcp14>MAY</bcp14> choose to only support contiguous SI values in an SFP. Such an implementation will not support receiving an SI value that is not present in the SFP and will discard the packets as described in <xref target="RFC8300"/>.</t>format="default"/>.</t> </section> </section> </section> <section anchor="selection"title="Selectionnumbered="true" toc="default"> <name>Selection within Service FunctionPaths">Paths</name> <t>As described in <xref target="overview"/>format="default"/>, the SPI/SI in the NSH passed back from an SFI to the SFF may leave the SFF with a choice ofnext hop SFTs,next-hop SFTs and a choice of SFIs for each SFT. That is, the SPI indicates an SFPR, and the SI indicates an entry in that SFPR. Each entry in an SFPR is a set of one or more SFT/SFIR-RD pairs. The SFFMUST<bcp14>MUST</bcp14> choose one of these, identify the SFF that supports the chosen SFI, and send the packet to thatnext hopnext-hop SFF.</t> <t>The choice be may offered for load balancing across multiple SFIs, or for discrimination between different actions necessary at a specific hop in the SFP. Different SFT values may exist at a given hop in an SFP to support several cases:<list style="symbols"> <t>There</t> <ul spacing="normal"> <li>There may be multiple instances of similar service functions that are distinguished by different SFT values. For example, firewalls made by vendor A and vendor B may need to be identified by different SFT values because, while they have similar functionality, their behavior is not identical. Then, some SFPs may limit the choice of SF at a given hop by specifying the SFT for vendor A, but other SFPs might not need to control whichvendor'svendor's SF is used and so can indicate that either SFT can beused.</t> <t>Thereused.</li> <li>There may be an obvious branch needed in anSFPSFP, such as the processing after a firewall where admitted packets continue along the SFP, but suspect packets are diverted to a "penalty box". In this case, the next hop in the SFP will be indicated with two different SFTvalues.</t> </list></t>values.</li> </ul> <t>In the typical case, the SFF chooses anext hopnext-hop SFF by looking at the set of all SFFs that support the SFs identified by the SI (that set having been advertised in individual SFIR advertisements), finding the one or more that are "nearest" in the underlay network, and choosing betweennext hopnext-hop SFFs using its own load-balancing algorithm.</t> <t>An SFI may influence this choice process by passing additional informationbackback, along with the packet and NSH. This information may influence local policy at the SFF to either cause it to favor anext hopnext-hop SFF (perhaps selecting one that is not nearest in theunderlay),underlay) ortoinfluence the load-balancing algorithm.</t> <t>This selection applies to the normalcase,case but also applies in the case of looping, jumping, and branching (see <xref target="looping"/>).</t>format="default"/>).</t> <t>Suppose an SFF in a particular service function overlay network (identified by a particular import RT, RT-z) needs to forward an NSH-encapsulated packet whose SPI is SPI-x and whose SI is SI-y. It does the following:<list style="numbers"> <t>It</t> <ol spacing="normal" type="1"><li>It looks for an installed SFPR that carries RT-z andthathas SPI-x in its NLRI. If there is none, then such packets cannot beforwarded.</t> <t>Fromforwarded.</li> <li>From the SFP attribute of that SFPR, it finds the Hop TLV with SI value set to SI-y. If there is no such Hop TLV, then such packets cannot beforwarded.</t>forwarded.</li> <li> <t>It then finds the "relevant" set of SFIRs by going through the list of SFT TLVs contained in the Hop TLV as follows:<list style="letters"> <t>An</t> <ol spacing="normal" type="A"><li>An SFIR is relevant if it carries RT-z, the SFT in its NLRI matches the SFT value in one of the SFT TLVs, and the RD value in its NLRI matches an entry in the list of SFIR-RDs in that SFTTLV.</t> <t>IfTLV.</li> <li>If an entry in the SFIR-RD list of an SFT TLV contains the value zero, then an SFIR is relevant if it carries RT-z and the SFT in its NLRI matches the SFT value in that SFT TLV.I.e.,That is, any SFIR in the service function overlay network defined by RT-z and with the correct SFT isrelevant.</t> <t>Ifrelevant.</li> <li>If a pool identifier is inuseuse, then an SFIR is relevant if it is a member of thepool.</t> </list></t> </list></t>pool.</li> </ol> </li> </ol> <t>Each of the relevant SFIRs identifies a singleSFI,SFI and contains aTunnel Encapsulationtunnel encapsulation attribute that specifies how to send a packet to that SFI. For a particular packet, the SFF chooses a particular SFI from the set of relevant SFIRs. This choice is made according to local policy.</t> <t>A typical policy might be to figure out the set of SFIs that areclosest,closest andtoload balance among them. But this is not the only possible policy.</t> <t>Thus, at any point in time when an SFF selects its next hop, it chooses from the intersection of the set ofnext hopnext-hop RDs contained in the SFPR and the RDs contained in theSFF'sSFF's local set of SFIRs (i.e., according to the determination of "relevance", above). If the intersection is null, the SFPR is unusable. Similarly, when this condition applies on theControllercontroller that originated the SFPR, itSHOULD<bcp14>SHOULD</bcp14> either withdraw the SFPR or re-advertise it with a new set of RDs for the affected hop.</t> </section> <section anchor="looping"title="Looping,numbered="true" toc="default"> <name>Looping, Jumping, andBranching">Branching</name> <t>As described in <xref target="overview"/>format="default"/>, an SFI or an SFF may cause a packet to "loop back" to a previous SF on a path in order that a sequence of functions may be re-executed. This is simply achieved by replacing the SI in the NSH with a highervaluevalue, instead of decreasing it as would normally be thecasecase, to determine the next hop in the path.</t> <t><xref target="overview"/>format="default"/> also describes how an SFI oranSFF may cause apacketspacket to "jump forward" to an SF on a path that is not the immediate next SF in the SFP. This is simply achieved by replacing the SI in the NSH with a lower value than would be achieved by decreasing it by the normal amount.</t> <t>A more complex option to move packets from one SFP to another is described in <xref target="RFC8300"/>format="default"/> and <xref target="overview"/>format="default"/>, where it is termed "branching". This mechanism allows an SFI or SFF to make a choice of downstream treatments for packets based on local policy and the output of the local SF. Branching is achieved by changing the SPI in the NSH to indicate the new path and setting the SI to indicate the point in the path at which the packets enter.</t> <t>Note that the NSH does not include a marker to indicate whether a specific packet has been around a loop before. Therefore, the use of NSH metadata(<xref<xref target="RFC8300"/>)format="default"/> may be required in order to prevent infinite loops.</t> <section anchor="changeseq"title="Protocolnumbered="true" toc="default"> <name>Protocol Control of Looping, Jumping, andBranching">Branching</name> <t>If the SFT value in an SFT TLV in an SFPR has theSpecial Purposespecial-purpose SFT value "Change Sequence" (see <xref target="iana"/>)format="default"/>), then this is an indication that the SFF may make a loop, jump, or branch according to local policy and information returned by the local SFI.</t> <t>In this case, the SPI and SI of the next hop are encoded in the eight bytes of an entry in the SFIR-RD list as follows:<list style="empty"> <t>3</t> <ul empty="true" spacing="normal"> <li>3 bytesSPI</t> <t>1 bytes SI</t> <t>4SPI</li> <li>1 byte SI</li> <li>4 bytes Reserved(SHOULD(<bcp14>SHOULD</bcp14> be set to zero andignored)</t> </list></t>ignored)</li> </ul> <t>If the SI in this encoding is not part of the SFPR indicated by the SPI in this encoding, then this is an explicit error thatSHOULD<bcp14>SHOULD</bcp14> be detected by the SFF when it parses the SFPR. The SFPRSHOULD NOT<bcp14>SHOULD NOT</bcp14> cause any forwarding state to be installed in theSFFSFF, and packets received with the SPI that indicates this SFPRSHOULD<bcp14>SHOULD</bcp14> be silently discarded.</t> <t>If the SPI in this encoding is unknown, the SFFSHOULD NOT<bcp14>SHOULD NOT</bcp14> install any forwarding state for thisSFPR,SFPR butMAY<bcp14>MAY</bcp14> hold the SFPR pending receipt of another SFPR that does use the encoded SPI.</t> <t>If the SPI matches the current SPI for the path, this is a loop or jump. In this case, if the SI is greater than or equal to the currentSISI, it is a loop. If the SPI matches and the SI is less than the next SI, it is a jump.</t> <t>If the SPI indicates another path, this is abranchbranch, and the SI indicates the point at which to enter that path.</t> <t>The Change Sequence SFT is just another SFT that may appear in a set of SFI/SFT tuples within an SI and is selected as described in <xref target="selection"/>.</t>format="default"/>.</t> <t>Note thatSpecial Purposespecial-purpose SFTsMUST NOT<bcp14>MUST NOT</bcp14> be advertised in SFIRs. If such an SFIR isreceivedreceived, itSHOULD<bcp14>SHOULD</bcp14> be ignored.</t> </section> <section anchor="implications"title="Implicationsnumbered="true" toc="default"> <name>Implications for ForwardingState">State</name> <t>Support for looping and jumping requires that the SFF has forwarding state established to an SFF that provides access to an instance of the appropriate SF. This means that the SFF must have seen the relevant SFIR advertisements and mush have known that it needed to create the forwarding state. This is a matter of local configuration andimplementation:implementation; for example, an implementation could be configured to install forwarding state for specific looping/jumping.</t> <t>Support for branching requires that the SFF has forwarding state established to an SFF that provides access to an instance of the appropriate entry SF on the other SFP. This means that the SFF must have seen the relevant SFIR and SFPR advertisements and known that it needed to create the forwarding state. This is a matter of local configuration andimplementation:implementation; for example, an implementation could be configured to install forwarding state for specific branching (identified by SPI and SI).</t> </section> </section> <section anchor="advanced"title="Advanced Topics">numbered="true" toc="default"> <name>Advanced Topics</name> <t>This section highlights several advanced topics introduced elsewhere in this document.</t> <section anchor="correlation"title="Correlatingnumbered="true" toc="default"> <name>Correlating Service Function PathInstances">Instances</name> <t>It is often useful to create bidirectional SFPs to enable packet flows to traverse the same set of SFs, but in the reverse order. However, packets on SFPs in the data plane (per <xref target="RFC8300"/>)format="default"/>) do not contain a direction indicator, so each direction must use a different SPI.</t> <t>As described in <xref target="assoctlv"/>format="default"/>, an SFPR can contain one or more correlators encoded in Association TLVs. If the Association Type indicates "BidirectionalSFP"SFP", then the SFP advertised in the SFPR is one direction of a bidirectional pair ofSFPsSFPs, where the other in the pair is advertised in the SFPR with RD as carried in theAssociated SFPR-RD"Associated SFPR-RD" field of the Association TLV. The SPI carried in theAssociated SPI"Associated SPI" field of the Association TLV provides a cross-check against the SPI advertised in the SFPR with RD as carried in theAssociated SFPR-RD"Associated SFPR-RD" field of the Association TLV.</t> <t>As noted in <xref target="assoctlv"/>,format="default"/>, when SFPRs reference each other, one SFPR advertisement will be received before the other. Therefore, processing of an association will require that the first SFPRisnot be rejected simply because the Associated SFPR-RD it carries is unknown. However, the SFP defined by the first SFPR is valid andSHOULD<bcp14>SHOULD</bcp14> be available for use as a unidirectionalSFPSFP, even in the absence of an advertisement of its partner.</t> <t>Furthermore, in error cases where SFPR-a associates with SFPR-b, but SFPR-b associates with SFPR-c such that a bidirectional pair of SFPs cannot be formed, the individual SFPs are still valid andSHOULD<bcp14>SHOULD</bcp14> be available for use as unidirectional SFPs. An implementationSHOULD<bcp14>SHOULD</bcp14> log thissituationsituation, because it represents aControllercontroller error.</t> <t>Usage of a bidirectional SFP may be programmed into theClassifiersclassifiers by theController.controller. Alternatively, aClassifierclassifier may look at incoming packets on a bidirectional packet flow, extract the SPI from the received NSH, and look up the SFPR to find thereverse directionreverse-direction SFP to use when it sends packets.</t> <t>See <xref target="example"/>format="default"/> for an example of how this works.</t> </section> <section anchor="stateful"title="Considerationsnumbered="true" toc="default"> <name>Considerations for Stateful ServiceFunctions">Functions</name> <t>Some service functions are stateful. That means that they build and maintain state derived from configuration orfromthe packet flows that they handle. In suchcasescases, it can be important or necessary that all packets from a flow continue to traverse the same instance of a service function so that the state can be leveraged and does not need to be regenerated.</t> <t>In the case of bidirectional SFPs, it may be necessary to traverse the same instances of a stateful service function in both directions. A firewall is a good example of such a service function.</t> <t>This issue becomes a concern where there are multiple parallel instances of a service function and a determination of which one to use could normally be left to the SFF as a load-balancing orlocal policylocal-policy choice.</t> <t>For theforward directionforward-direction SFP, the concern is that the same choice ofservice functionSF is made for all packets of a flow under normal network conditions. It may be possible to guarantee that theload balancingload-balancing functions applied in the SFFs are stable and repeatable, but aControllercontroller that constructs SFPs might not want to trust to this. TheControllercontroller can, in these cases, build a number of more specificSFPsSFPs, each traversing a specific instance of the stateful SFs. In this case, theload balancingload-balancing choice can be left up to theClassifier. Thusclassifier. Thus, theClassifierclassifier selects which instance of a stateful SF is used by a particular flow by selecting the SFP that the flow uses.</t> <t>For bidirectional SFPs where the same instance of a stateful SF must be traversed in both directions, it is not enough to leave the choice ofservice function instanceSFI as a localchoicechoice, even if the load balancing isstablestable, because coordination would be required between the decision points in the forward and reversedirectionsdirections, and this may be hard to achieve in all cases except where it is the same SFF that makes the choice in both directions.</t> <t>Note that this approach necessarily increases the amount of SFP state in the network (i.e., there are more SFPs). It is possible to mitigate this effect by careful construction of SFPs built from a concatenation of other SFPs.</t> <t><xref target="examplestate"/>format="default"/> includes some simple examples of SFPs for statefulservice functions.</t>SFs.</t> </section> <section anchor="private"title="VPNnumbered="true" toc="default"> <name>VPN Considerations and Private ServiceFunctions">Functions</name> <t>Likely deployments include reserving specific instances ofService FunctionsSFs for specific customers or allowing customers to deploy their ownService FunctionsSFs within the network. BuildingService FunctionsSFs in such environments requires that suitable identifiersarebe used to ensure that SFFs distinguish which SFIs can be used and which cannot.</t> <t>This problem is similar tohowa problem in the way that VPNs are supported and is solved in a similar way. TheRT"RT" field is used to indicate a set ofService FunctionsSFs from which all choices must be made.</t> </section> <section anchor="fspecclassy"title="Flownumbered="true" toc="default"> <name>Flow Specification for SFCClassifiers">Classifiers</name> <t><xreftarget="I-D.ietf-idr-rfc5575bis" />target="RFC8955" format="default"/> defines a set of BGP routes that can be used to identify the packets in a given flow using fields in the header of each packet, and a set ofactions,actions -- encoded asextended communities,Extended Communities -- that can be used to disposition those packets. This document enables the use of these mechanisms by SFCClassifiersclassifiers by defining a new actionextended communityExtended Community called "Flow Specification for SFCClassifiers"Classifiers", identified by the valueTBD4.0x0d. Note that implementation of this section of this specification will beControllerscontrollers orClassifiersclassifiers communicating with each other directly for the purpose of instructing theClassifierclassifier how to place packets onto an SFP.In orderSo that the implementation ofClassifiersclassifiers can be keptsimplesimple, and to avoid the confusion between thepurposepurposes of differentextended communities,Extended Communities, aController MUST NOTcontroller <bcp14>MUST NOT</bcp14> include other actionextended communitiesExtended Communities at the same time as a "Flow Specification for SFC Classifiers"extended community: aExtended Community. A "Flow Specification for SFC Classifiers" Traffic Filtering Action Extended Community advertised with any other Traffic Filtering Action Extended CommunityMUST<bcp14>MUST</bcp14> be treated as malformed in line with <xreftarget="I-D.ietf-idr-rfc5575bis" />target="RFC8955" format="default"/> and result in theFlow Specificationflow-specification UPDATE message being handled astreat-as-withdraw"treat-as-withdraw", according to <xref target="RFC7606"/> Section 2.</t>section="2" sectionFormat="comma"/>.</t> <t>To put theFlow Specificationflow specification intocontextcontext, when multipleSFCservice function chaining overlays are present in one network, each FlowSpec updateMUST<bcp14>MUST</bcp14> be tagged with the route target of the overlay or VPN network for which it is intended.</t> <t>Thisextended communityExtended Community is encoded as an 8-octet value, as shown in <xreftarget="fspecclassyfig"/>.</t>target="fspecclassyfig" format="default"/>.</t> <figureanchor="fspecclassyfig" title="Theanchor="fspecclassyfig"> <name>The Format of the Flow Specification for SFC Classifiers ExtendedCommunity"> <artwork> <![CDATA[Community</name> <artwork name="" type="" align="left" alt=""><![CDATA[ 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type=0x80 |Sub-Type=TBD4Sub-Type=0x0d | SPI | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SPI (cont.) | SI | SFT | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+]]> </artwork>]]></artwork> </figure> <t>Theextended communityExtended Community contains the Service Path Identifier (SPI), Service Index (SI), and Service Function Type(SFT)(SFT), as defined elsewhere in this document. Thus, each action extended community defines the entry point (not necessarily the first hop) into a specificservice function path.SFP. This allows, for example, different flows to enter the sameservice function pathSFP at different points.</t> <t>Notethatthat, according to <xreftarget="I-D.ietf-idr-rfc5575bis" />target="RFC8955" format="default"/>, a givenFlow Specificationflow-specification update may include multiple of these actionextended communities.Extended Communities. If a given action extended community does not contain an installed SFPR with the specified {SPI, SI,SFT}SFT}, itMUST NOT<bcp14>MUST NOT</bcp14> be used for dispositioning the packets of the specified flow.</t> <t>The normal case of packet classification forSFCservice function chaining will see a packet enter the SFP at its first hop. In thiscasecase, the SI in theextended communityExtended Community issuperfluoussuperfluous, and the SFT may also be unnecessary. To allow these cases to be handled, a special meaning is assigned toa Service Indexan SI of zero (not a valid value) and an SFT of zero (a reserved value in the registry--- see <xref target="SFTreg"/>). <list style="symbols"> <t>Ifformat="default"/>). </t> <ul spacing="normal"> <li>If an SFC Classifiers Extended Community is received with SI =00, then it means that the first hop of the SFP indicated by the SPIMUST<bcp14>MUST</bcp14> beused.</t>used.</li> <li> <t>If an SFC Classifiers Extended Community is received with SFT =00, then there are twosub-cases: <list style="symbols"> <t>Ifsubcases: </t> <ul spacing="normal"> <li>If there is a choice of SFT in the hop indicated by the value of the SI (including SI =0)0), then SFT = 0 means there is a free choiceaccording to local policyof which SFT touse).</t> <t>Ifuse, according to local policy).</li> <li>If there is no choice of SFT in the hop indicated by the value of SI, then SFT = 0 means that the value of the SFT at thathophop, as indicated in the SFPR for the indicatedSPI MUSTSPI, <bcp14>MUST</bcp14> beused.</t> </list></t> </list></t>used.</li> </ul> </li> </ul> <t>One of the filters that theFlow Specificationflow specification may describe is the VPN to which the traffic belongs. Additionally, as noted above, to put the indicated SPI into context when multiple SFC overlays are present in one network, each FlowSpec updateMUST<bcp14>MUST</bcp14> be tagged with the route target of the overlay or VPN network for which it is intended.</t> <t>Note that future extensions might be made to the Flow Specification for SFC Classifiers Extended Community to provide instruction to theClassifierclassifier about what metadata to add to packets that it classifies for forwarding on a specificSFP, butSFP; however, that is outside the scope of this document.</t> </section> <section anchor="representation"title="Choicenumbered="true" toc="default"> <name>Choice of Data Plane SPI/SIRepresentation">Representation</name> <t>This document ties together the control and data planes ofan SFCa service function chaining overlay network through the use of the SPI/SIwhichthat is nominally carried in the NSH of a given packet. However, in order to handle situations in which the NSH is not ubiquitously deployed, it is also possible to use alternative data plane representations of the SPI/SI by carrying the identical semantics in other protocolfieldsfields, such as MPLS labels <xref target="RFC8595"/>.</t>format="default"/>.</t> <t>This document defines a newsub-TLVSub-TLV for theTunnel Encapsulationtunnel encapsulation attribute <xreftarget="I-D.ietf-idr-tunnel-encaps" />,target="RFC9012" format="default"/>, the SPI/SI Representationsub-TLVSub-TLV of typeTBD5.16. Thissub-TLV MAYSub-TLV <bcp14>MAY</bcp14> be present in each Tunnel TLV contained in aTunnel Encapsulationtunnel encapsulation attribute when the attribute is carried by an SFIR. Thevalue"Value" field of thissub-TLVSub-TLV is atwo octettwo-octet field of flags numbered counting from thethemost significant bit, each of which describes how the originating SFF expects to see the SPI/SI represented in the data plane for packets carried in the tunnels described by the Tunnel TLV.</t> <t>The following bits are defined by this document and are tracked in an IANA registry described in <xref target="IANAbits"/>: <list style="hanging"> <t hangText="Bit TBD9:">Ifformat="default"/>: </t> <dl newline="false" spacing="normal"> <dt>Bit 0:</dt> <dd>If this bit issetset, the NSH is to be used to carry the SPI/SI in the dataplane.</t> <t hangText="Bit TBD10:">Ifplane.</dd> <dt>Bit 1:</dt> <dd>If this bit issetset, two labels in an MPLS label stack are to be used as described in <xref target="MPLS-NSH"/>.</t> </list></t>format="default"/>.</dd> </dl> <t>If a given Tunnel TLV does not contain an SPI/SI Representationsub-TLVSub-TLV, then itMUST<bcp14>MUST</bcp14> be processed as if such asub-TLVSub-TLV is present with BitTBD90 set and no other bits set. That is, the absence of thesub-TLV SHALLSub-TLV <bcp14>SHALL</bcp14> be interpreted to mean that the NSH is to be used.</t> <t>If a given Tunnel TLV contains an SPI/SI Representationsub-TLVSub-TLV withvaluea "Value" field that has no flagsetset, then the tunnel indicated by the Tunnel TLVMUST NOT<bcp14>MUST NOT</bcp14> be used for forwarding SFC packets. If a given Tunnel TLV contains an SPI/SI Representationsub-TLVSub-TLV with both bitTBD90 and bitTBD10 set1 set, then the tunnel indicated by the Tunnel TLVMUST NOT<bcp14>MUST NOT</bcp14> be used for forwarding SFC packets. The meaning and rules for the presence of other bits is to be defined in future documents, but implementations of this specificationMUST<bcp14>MUST</bcp14> set other bits to zero and ignore them on receipt.</t> <t>If a given Tunnel TLV contains more than one SPI/SI Representationsub-TLVSub-TLV, then the first oneMUST<bcp14>MUST</bcp14> be considered and subsequent instancesMUST<bcp14>MUST</bcp14> be ignored.</t> <t>Note that the MPLS representation of the logical NSH may be used even if the tunnel is not an MPLS tunnel. Conversely, MPLS tunnels may be used to carry other encodings of the logical NSH (specifically, the NSH itself). It is a requirement that both ends of a tunnel over the underlay network know that the tunnel is used forSFCservice function chaining and know what form of NSH representation is used. The signaling mechanism described here allows coordination of this information.</t> <section anchor="MPLS-NSH"title="MPLSnumbered="true" toc="default"> <name>MPLS Representation of theSPI/SI">SPI/SI</name> <t>If bitTBD101 is set in thein theSPI/SI Representationsub-TLVSub-TLV, then labels in the MPLS label stack are used to indicate SFC forwarding and processing instructions to achieve the semantics of a logical NSH. The label stack is encoded as shown in <xref target="RFC8595"/>.</t>format="default"/>.</t> </section> </section> <section anchor="swapOp"title="MPLSnumbered="true" toc="default"> <name>MPLS Label Swapping/StackingOperation">Operation</name> <t>When aClassifierclassifier constructs an MPLS label stack for anSFPSFP, it starts with thatSFP'sSFP's last hop. If the last hop requires an {SPI, SI} label pair for label swapping, it pushes the SI (set to the SI value of the last hop) and theSFP'sSFP's SPI onto the MPLS label stack. If the last hop requires a {context label, SFI label} label pair for labelstackingstacking, it selects a specific SFIR and pushes thatSFIR'sSFIR's SFI label and context label onto the MPLS label stack.</t> <t>TheClassifierclassifier then moves sequentially back through the SFP one hop at a time. For each hop, if the hop requires an {SPI,SI]}SI} and there is an {SPI, SI} at the top of the MPLS label stack, the SI is set to the SI value of the current hop. If there is not an {SPI, SI} at the top of the MPLS label stack, it pushes the SI (set to the SI value of the current hop) and theSFP'sSFP's SPI onto the MPLS label stack.</t> <t>If the hop requires a {context label, SFI label}, it selects a specific SFIR and pushes thatSFIR'sSFIR's SFI label and context label onto the MPLS label stack.</t> </section> <section anchor="mpls-encaps"title="Supportnumbered="true" toc="default"> <name>Support for MPLS-Encapsulated NSHPackets">Packets</name> <t><xref target="RFC8596"/>format="default"/> describes how to transport SFC packets using the NSH over an MPLS transport network. SignalingMPLS encapsulation of SFC packets using the NSHthat this approach is in use isalsosupported by this documentby using theas follows:</t> <ul spacing="normal"> <li>A "BGP Tunnel EncapsulationAttribute Sub-TLV"Attribute" Sub-TLV is included with the codepoint 10 (representing "MPLS Label Stack") from the "BGP Tunnel Encapsulation Attribute Sub-TLVs" registry defined in <xreftarget="I-D.ietf-idr-tunnel-encaps" />, and also using thetarget="RFC9012" format="default"/>.</li> <li>An "SFP Traversal With MPLS LabelStack TLV" and theStack" TLV is included containing an "SPI/SIRepresentation sub-TLV"Representation" Sub-TLV with bitTBD90 set and bitTBD10 cleared.</t>1 cleared.</li></ul> <t>In thiscasecase, the MPLS label stack constructed by the SFF to forward a packet to the next SFF on the SFP will consist of the labels needed to reach that SFF, and if label stacking isusedused, it will also include the labels advertised in the MPLS Label Stacksub-TLVSub-TLV and the labels remaining in the stack needed to traverse the remainder of the SFP.</t> </section> </section> <section anchor="example"title="Examples">numbered="true" toc="default"> <name>Examples</name> <t>Most of the examples in this section use IPv4 addressing. But there is nothing special about IPv4 in the mechanisms described in this document, and they are equally applicable to IPv6. A few examples using IPv6 addressing are provided in <xref target="v6samples"/>.</t>format="default"/>.</t> <t>Assume we have a service function overlay network with four SFFs (SFF1,SFF3,SFF2, SFF3, and SFF4). The SFFs have addresses in the underlay network as follows:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFF1 192.0.2.1 SFF2 192.0.2.2 SFF3 192.0.2.3 SFF4 192.0.2.4]]> </artwork> </figure>]]></sourcecode> <t>Each SFF provides access to some SFIs from the fourService Function TypesSFTs, SFT=41, SFT=42, SFT=43, andSFT=44SFT=44, as follows:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFF1 SFT=41 and SFT=42 SFF2 SFT=41 and SFT=43 SFF3 SFT=42 and SFT=44 SFF4 SFT=43 and SFT=44]]> </artwork> </figure>]]></sourcecode> <t>The service function network also contains aControllercontroller with address 198.51.100.1.</t> <t>This example service function overlay network is shown in <xref target="examplefig"/>.</t>format="default"/>.</t> <figureanchor="examplefig" title="Exampleanchor="examplefig"> <name>Example Service Function OverlayNetwork"> <artwork> <![CDATA[Network</name> <artwork name="" type="" align="left" alt=""><![CDATA[ -------------- | Controller | | 198.51.100.1 | ------ ------ ------ ------ -------------- | SFI | | SFI | | SFI | | SFI | |SFT=41| |SFT=42| |SFT=41| |SFT=43| ------ ------ ------ ------ \ / \ / --------- --------- ---------- | SFF1 | | SFF2 | Packet --> | | |192.0.2.1| |192.0.2.2| Flows --> |Classifier| --------- --------- -->Dest | | --> ---------- --------- --------- | SFF3 | | SFF4 | |192.0.2.3| |192.0.2.4| --------- --------- / \ / \ ------ ------ ------ ------ | SFI | | SFI | | SFI | | SFI | |SFT=42| |SFT=44| |SFT=43| |SFT=44| ------ ------ ------ ------]]> </artwork>]]></artwork> </figure> <t>The SFFs advertise routes to the SFIs they support. These advertisements containRoute DistinguishersRDs that are set according to the networkoperator'soperator's configuration model. In all of these IPv4examplesexamples, we use RDs oftypeType 1 such that the available six octets are partitioned as four octets for the IPv4 address of the advertising SFF, and two octets that are a local index of the SFI. This scheme is chosen purely for convenience of documentation, and an operator is totally free to use any other scheme so long as it conforms to the definitions of SFIR and SFPR in Sections <xref target="sfiRoutes"/>format="counter"/> and <xref target="sfpRoutes"/>.</t> <t>Thusformat="counter"/>.</t> <t>Thus, we see the following SFIRs advertised:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ RD = 192.0.2.1/1, SFT = 41 RD = 192.0.2.1/2, SFT = 42 RD = 192.0.2.2/1, SFT = 41 RD = 192.0.2.2/2, SFT = 43 RD = 192.0.2.3/7, SFT = 42 RD = 192.0.2.3/8, SFT = 44 RD = 192.0.2.4/5, SFT = 43 RD = 192.0.2.4/6, SFT = 44]]> </artwork> </figure>]]></sourcecode> <t>Note that the addressing used for communicating between SFFs is taken from theTunnel Encapsulationtunnel encapsulation attribute of the SFIR and not from the SFIR-RD.</t> <section anchor="exampleexplicit"title="Examplenumbered="true" toc="default"> <name>Example Explicit SFPWithwith NoChoices" >Choices</name> <t>Consider the following SFPR.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP1: RD = 198.51.100.1/101, SPI = 15, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, RD = 192.0.2.2/2]]]> </artwork> </figure>]]></sourcecode> <t>TheService Function PathSFP consists of an SF oftypeType 41 located atSFF1SFF1, followed by an SF oftypeType 43 located at SFF2. This path is fullyexplicitexplicit, and each SFF is offered no choice in forwarding packets along the path.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (15). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 has no flexibility in the choice of SFF to support thenext hopnext-hop SFI and will forward the packet toSFF2SFF2, which will send the packets to the SFI that supports SFT 43 before forwarding the packets to their destinations.</t> </section> <section anchor="examplechoice"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Choice ofSFIs" > <figure> <artwork> <![CDATA[SFIs</name> <sourcecode><![CDATA[ SFP2: RD = 198.51.100.1/102, SPI = 16, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, {RD = 192.0.2.2/2, RD = 192.0.2.4/5 } ]]]> </artwork> </figure>]]></sourcecode> <t>In thisexampleexample, the path also consists of an SF oftypeType 41 located atSFF1SFF1, and this is followed by an SF oftype 43, butType 43. However, in thiscasecase, the SI = 250 contains a choice between the SFI located at SFF2 and the SFI located at SFF4.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (16). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a choice ofnext hop SFFnext-hop SFFs to execute the next hop in the path. It can either forward packets to SFF2 or SFF4 to execute a function oftypeType 43. It uses its localload balancingload-balancing algorithm to make this choice. The chosen SFF will send the packets to the SFI that supports SFT 43 before forwarding the packets to their destinations.</t> </section> <section anchor="exampleopen"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Open Choice ofSFIs" > <figure> <artwork> <![CDATA[SFIs</name> <sourcecode><![CDATA[ SFP3: RD = 198.51.100.1/103, SPI = 17, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 44, RD = 0]]]> </artwork> </figure>]]></sourcecode> <t>In thisexampleexample, the path also consists of an SF oftypeType 41 located atSFF1SFF1, and this is followed by an SI with an RD of zero and SF oftypeType 44. This means that a choice can be made between any SFF that supports an SFI oftypeType 44.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (17). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a free choice ofnext hop SFFnext-hop SFFs to execute the next hop in thepathpath, selecting between all SFFs that support SFs oftypeType 44. Looking at the SFIRs it has received, SFF1 knows that SFtypeType 44 is supported by SFF3 and SFF4. SFF1 uses its localload balancingload-balancing algorithm to make this choice. The chosen SFF will send the packets to the SFI that supports SFT 44 before forwarding the packets to their destinations.</t> </section> <section anchor="examplesft"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Choice ofSFTs" > <figure> <artwork> <![CDATA[SFTs</name> <sourcecode><![CDATA[ SFP4: RD = 198.51.100.1/104, SPI = 18, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, {SFT = 43, RD = 192.0.2.2/2, SFT = 44, RD = 192.0.2.3/8 } ]]]> </artwork> </figure>]]></sourcecode> <t>This example provides a choice of SF type in the second hop in the path. The SI of 250 indicates a choice between SFtypeType 43 located at SF2 and SFtypeType 44 located at SF3.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (18). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a free choice ofnext hop SFFnext-hop SFFs to execute the next hop in thepathpath, selecting between all SFFs that support an SF oftypeType 43 andSFF3 thatSFF3, which supports an SF oftypeType 44. These may be completely different functions that are to be executed dependent on specific conditions, or they may be similar functions identified with different type identifiers (such as firewalls from different vendors). SFF1 uses its local policy andload balancingload-balancing algorithm to make thischoice,choice and may use additional information passed back from the local SFI to help inform its selection. The chosen SFF will send the packets to the SFI that supports thechosechosen SFT before forwarding the packets to their destinations.</t> </section> <section anchor="exampleco"title="Examplenumbered="true" toc="default"> <name>Example Correlated BidirectionalSFPs" > <figure> <artwork> <![CDATA[SFPs</name> <sourcecode><![CDATA[ SFP5: RD = 198.51.100.1/105, SPI = 19, Assoc-Type = 1, Assoc-RD = 198.51.100.1/106, Assoc-SPI = 20, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, RD = 192.0.2.2/2] SFP6: RD = 198.51.100.1/106, SPI = 20, Assoc-Type = 1, Assoc-RD = 198.51.100.1/105, Assoc-SPI = 19, [SI = 254, SFT = 43, RD = 192.0.2.2/2], [SI = 249, SFT = 41, RD = 192.0.2.1/1]]]> </artwork> </figure>]]></sourcecode> <t>This example demonstrates correlation of two SFPs to form a bidirectionalSFPSFP, as described in <xref target="correlation"/>.</t>format="default"/>.</t> <t>Two SFPRs are advertised by theController.controller. They have different SPIs (19 and20)20), so they are known to be separate SFPs, but they both have Association TLVs with Association Type set to11, indicating bidirectional SFPs. Each has anAssociated SFPR-RD"Associated SFPR-RD" field containing the value of the other SFPR-RD tocorrelatedcorrelate the two SFPs as a bidirectional pair.</t> <t>As can be seen from the SFPRs in this example, the paths are symmetric: the hops in SFP5 appear in the reverse order in SFP6.</t> </section> <section anchor="exampleass"title="Examplenumbered="true" toc="default"> <name>Example Correlated Asymmetrical BidirectionalSFPs" > <figure> <artwork> <![CDATA[SFPs</name> <sourcecode><![CDATA[ SFP7: RD = 198.51.100.1/107, SPI = 21, Assoc-Type = 1, Assoc-RD = 198.51.100.1/108, Assoc-SPI = 22, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, RD = 192.0.2.2/2] SFP8: RD = 198.51.100.1/108, SPI = 22, Assoc-Type = 1, Assoc-RD = 198.51.100.1/107, Assoc-SPI = 21, [SI = 254, SFT = 44, RD = 192.0.2.4/6], [SI = 249, SFT = 41, RD = 192.0.2.1/1]]]> </artwork> </figure>]]></sourcecode> <t>Asymmetric bidirectional SFPs can also be created. This example shows a pair of SFPs with distinct SPIs (21 and 22) that are correlated in the same way as in the example in <xref target="exampleco"/>.</t>format="default"/>.</t> <t>However, unlike in that example, the SFPs are different in each direction. Both paths include a hop of SFtypeType 41, but SFP7 includes a hop of SFtypeType 43 supported atSFF2SFF2, while SFP8 includes a hop of SFtypeType 44 supported at SFF4.</t> </section> <section anchor="exampleloop"title="Examplenumbered="true" toc="default"> <name>Example Looping in anSFP" > <figure> <artwork> <![CDATA[SFP</name> <sourcecode><![CDATA[ SFP9: RD = 198.51.100.1/109, SPI = 23, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 44, RD = 192.0.2.4/5], [SI = 245, {SFT = 1, RD = {SPI=23, SI=255, Rsv=0}, SFT = 42, RD = 192.0.2.3/7 } ]]]> </artwork> </figure>]]></sourcecode> <t>Looping and jumping are described in <xref target="looping"/>.format="default"/>. This example shows an SFP that contains an explicit loop-back instruction that is presented as a choice within an SFP hop.</t> <t>The first two hops in the path (SI = 255 and SI = 250) are normal. That is, the packets will be delivered to SFF1 and SFF4 in turn for execution of SFs oftypeType 41 and4444, respectively.</t> <t>The third hop (SI = 245) presents SFF4 with a choice of next hop. It can either forward the packets to SFF3 for an SF oftypeType 42 (the secondchoice),choice) or it can loop back.</t> <t>The loop-back entry in the SFPR for SI = 245 is indicated by thespecial purposespecial-purpose SFT value 1 ("Change Sequence"). Within this hop, the RD is interpreted as encoding the SPI and SI of the next hop (see <xref target="changeseq"/>.format="default"/>). In thiscasecase, the SPI is2323, which indicates that this is a loop orbranch:branch, i.e., the next hop is on the same SFP. The SI is set to255:255; this is a higher number than the current SI(245)(245), indicating a loop.</t> <t>SFF4 must make a choice between these two next hops.Either the packetsThe packet will be either forwarded to SFF3 with the NSH SI decreased to 245 or looped back to SFF1 with the NSH SI reset to 255. This choice will be made according to local policy, information passed back by the local SFI, and details in thepackets'packets' metadata that are used to prevent infinite looping.</t> </section> <section anchor="examplebranch"title="Examplenumbered="true" toc="default"> <name>Example Branching in anSFP" > <figure> <artwork> <![CDATA[SFP</name> <sourcecode><![CDATA[ SFP10: RD = 198.51.100.1/110, SPI = 24, [SI = 254, SFT = 42, RD = 192.0.2.3/7], [SI = 249, SFT = 43, RD = 192.0.2.2/2] SFP11: RD = 198.51.100.1/111, SPI = 25, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 1, RD = {SPI=24, SI=254, Rsv=0}]]]> </artwork> </figure>]]></sourcecode> <t>Branching follows a similar procedure to that for looping (andjumping)jumping), as shown in <xref target="exampleloop"/> howeverformat="default"/>. However, there are two SFPs involved.</t> <t>SFP10 shows a normal path with packets forwarded to SFF3 and SFF2 for execution of service functions oftypeType 42 and4343, respectively.</t> <t>SFP11 starts as normal (SFF1 for an SF oftypeType 41), but then SFF1 processes the next hop in the path and finds a "Change Sequence"Special Purposespecial-purpose SFT. TheSFIR-RD"SFIR-RD" field includes an SPI of2424, which indicates SFP10, not the current SFP. The SI in the SFIR-RD is 254, so SFF1 knows that it must set the SPI/SI in the NSH to 24/254 and send the packets to the appropriateSFFSFF, as advertised in the SFPR for SFP10 (that is, SFF3).</t> </section> <section anchor="examplestate"title="Examplesnumbered="true" toc="default"> <name>Examples of SFPs with Stateful ServiceFunctions" >Functions</name> <t>This section provides some examples to demonstrate establishing SFPs when there is a choice of service functions at a particular hop, and where consistency of choice is required in both directions. The scenarios that give rise to this requirement are discussed in <xref target="stateful"/>.</t>format="default"/>.</t> <section anchor="stateegsff"title="Forwardnumbered="true" toc="default"> <name>Forward and Reverse Choice Made at theSFF" >SFF</name> <t>Consider the topology shown in <xreftarget="egsfffig"/>.target="egsfffig" format="default"/>. There are three SFFs arranged neatly in a line, and the middle one (SFF2) supports three SFIs all of SFT 42. These three instances can be used by SFF2 to load balance so that no one instance is swamped.</t> <figureanchor="egsfffig" title="Exampleanchor="egsfffig"> <name>Example Where ChoiceisIs Made at theSFF"> <artwork> <![CDATA[SFF</name> <artwork name="" type="" align="left" alt=""><![CDATA[ ------ ------ ------ ------ ------ | SFI | | SFIa | | SFIb | | SFIc | | SFI | |SFT=41| |SFT=42| |SFT=42| |SFT=42| |SFT=43| ------ ------\ ------ /------ ------ \ \ | / / --------- --------- --------- ---------- | SFF1 | | SFF2 | | SFF3 | --> | |..|192.0.2.1|...|192.0.2.2|...|192.0.2.3|--> --> |Classifier| --------- --------- --------- | | ----------]]> </artwork>]]></artwork> </figure> <t>This leads to the following SFIRs being advertised.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ RD = 192.0.2.1/11, SFT = 41 RD = 192.0.2.2/11, SFT = 42 (for SFIa) RD = 192.0.2.2/12, SFT = 42 (for SFIb) RD = 192.0.2.2/13, SFT = 42 (for SFIc) RD = 192.0.2.3/11, SFT = 43]]> </artwork> </figure>]]></sourcecode> <t>The controller can create a single forward SFP(SFP12)(SFP12), giving SFF2 the choice of which SFI to use to provide a function of SFT4242, as follows. The load-balancing choice between the three available SFIs is assumed to be within the capabilities of theSFFSFF, and if the SFs arestatefulstateful, it is assumed that the SFF knows this and arranges load balancing in a stable, flow-dependent way.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP12: RD = 198.51.100.1/112, SPI = 26, Assoc-Type = 1, Assoc-RD = 198.51.100.1/113, Assoc-SPI = 27, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, {RD = 192.0.2.2/11, 192.0.2.2/12, 192.0.2.2/13 }], [SI = 253, SFT = 43, RD = 192.0.2.3/11]]]> </artwork> </figure>]]></sourcecode> <t>The reverse SFP (SFP13) in this case may also be created as shownbelowbelow, using association with the forward SFP and giving the load-balancing choice to SFF2. This is safe, even in the case that the SFs oftypeType 42 arestatefulstateful, because SFF2 is doing the load balancing in both directions and can apply the same algorithm to ensure that packets associated with the same flow use the same SFI regardless of the direction of travel.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP13: RD = 198.51.100.1/113, SPI = 27, Assoc-Type = 1, Assoc-RD = 198.51.100.1/112, Assoc-SPI = 26, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, {RD = 192.0.2.2/11, 192.0.2.2/12, 192.0.2.2/13 }], [SI = 253, SFT = 41, RD = 192.0.2.1/11]]]> </artwork> </figure>]]></sourcecode> <t>How an SFF knows that an attached SFI is stateful is out of the scope of this document. It is assumed that this will form part of the process by which SFIs are registered as local to SFFs. <xref target="stateful"/>format="default"/> provides additional observations about the coordination of the use of stateful SFIs in the case of bidirectional SFPs.</t> <t>In general, the problems of load balancing and the selection of the same SFIs in both directions of a bidirectional SFP can be addressed by using sufficiently precisely specified SFPs (specifying the exact SFIs to use) and suitable programming of theClassifiersclassifiers at each end of the SFPs to make sure that the matching pair of SFPs are used.</t> </section> <section anchor="stateeg1pll"title="Parallelnumbered="true" toc="default"> <name>Parallel End-to-End SFPs with SharedSFF" >SFF</name> <t>The mechanism described in <xref target="stateegsff"/>format="default"/> might not be desirable because of the functional assumptions it places on SFF2 to be able to load balance with suitable flow identification, stability, and equality in both directions. Instead, it may be desirable to place the responsibility for flow classification in theClassifierclassifier and let it determine load balancing with the implied choice of SFIs.</t> <t>Consider the network graph as shown in <xref target="egsfffig"/>format="default"/> and with the same set of SFIRs as listed in <xref target="stateegsff"/>.format="default"/>. In thiscasecase, the controller could specify three forward SFPs with their corresponding associated reverse SFPs. Each bidirectional pair of SFPs uses a different SFI for the SF oftypeType 42. The controller can instruct theClassifierclassifier how to place traffic on the three bidirectional SFPs, or it can treat them as agroupgroup, leaving theClassifierclassifier responsible for balancing the load.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP14: RD = 198.51.100.1/114, SPI = 28, Assoc-Type = 1, Assoc-RD = 198.51.100.1/117, Assoc-SPI = 31, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.2/11], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP15: RD = 198.51.100.1/115, SPI = 29, Assoc-Type = 1, Assoc-RD = 198.51.100.1/118, Assoc-SPI = 32, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.2/12], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP16: RD = 198.51.100.1/116, SPI = 30, Assoc-Type = 1, Assoc-RD = 198.51.100.1/119, Assoc-SPI = 33, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.2/13], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP17: RD = 198.51.100.1/117, SPI = 31, Assoc-Type = 1, Assoc-RD = 198.51.100.1/114, Assoc-SPI = 28, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.2/11], [SI = 253, SFT = 41, RD = 192.0.2.1/11] SFP18: RD = 198.51.100.1/118, SPI = 32, Assoc-Type = 1, Assoc-RD = 198.51.100.1/115, Assoc-SPI = 29, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.2/12], [SI = 253, SFT = 41, RD = 192.0.2.1/11] SFP19: RD = 198.51.100.1/119, SPI = 33, Assoc-Type = 1, Assoc-RD = 198.51.100.1/116, Assoc-SPI = 30, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.2/13], [SI = 253, SFT = 41, RD = 192.0.2.1/11]]]> </artwork> </figure>]]></sourcecode> </section> <section anchor="stateeg2pll"title="Parallelnumbered="true" toc="default"> <name>Parallel End-to-End SFPs with SeparateSFFs" >SFFs</name> <t>While the examples in Sections <xref target="stateegsff"/>format="counter"/> and <xref target="stateeg1pll"/>format="counter"/> place the choice of SFI as subtended from the same SFF, it is also possible that the SFIs are each subtended from a differentSFFSFF, as shown in <xref target="eg2pllfig"/>.format="default"/>. In thiscasecase, it is harder to coordinate the choices for forward and reverse paths without some form of coordination between SFF1 and SFF3.ThereforeTherefore, it would be normal to consider end-to-end parallelSFPsSFPs, as described in <xref target="stateeg1pll"/>.</t>format="default"/>.</t> <figureanchor="eg2pllfig" title="Secondanchor="eg2pllfig"> <name>Second ExampleWithwith Parallel End-to-EndSFPs"> <artwork> <![CDATA[SFPs</name> <artwork name="" type="" align="left" alt=""><![CDATA[ ------ | SFIa | |SFT=42| ------ ------ | | SFI | --------- |SFT=41| | SFF5 | ------ ..|192.0.2.5|.. | ..: --------- :.. ---------.: :.--------- ---------- | SFF1 | --------- | SFF3 | --> | |..|192.0.2.1|....| SFF6 |....|192.0.2.3| --> --> |Classifier| ---------: |192.0.2.6| :--------- | | : --------- : | ---------- : | : ------ : ------ : | SFI | :.. | SFIb | ..: |SFT=43| :.. |SFT=42| ..: ------ : ------ : :.---------.: | SFF7 | |192.0.2.7| --------- | ------ | SFIc | |SFT=42| ------]]> </artwork>]]></artwork> </figure> <t>In this case, five SFIRs are advertised as follows:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ RD = 192.0.2.1/11, SFT = 41 RD = 192.0.2.5/11, SFT = 42 (for SFIa) RD = 192.0.2.6/11, SFT = 42 (for SFIb) RD = 192.0.2.7/11, SFT = 42 (for SFIc) RD = 192.0.2.3/11, SFT = 43]]> </artwork> </figure>]]></sourcecode> <t>In thiscasecase, the controller could specify three forward SFPs with their corresponding associated reverse SFPs. Each bidirectional pair of SFPs uses a different SFF and SFI for the middle hop (for an SF oftypeType 42). The controller can instruct theClassifierclassifier how to place traffic on the three bidirectional SFPs, or it can treat them as agroupgroup, leaving theClassifierclassifier responsible for balancing the load.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP20: RD = 198.51.100.1/120, SPI = 34, Assoc-Type = 1, Assoc-RD = 198.51.100.1/123, Assoc-SPI = 37, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.5/11], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP21: RD = 198.51.100.1/121, SPI = 35, Assoc-Type = 1, Assoc-RD = 198.51.100.1/124, Assoc-SPI = 38, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.6/11], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP22: RD = 198.51.100.1/122, SPI = 36, Assoc-Type = 1, Assoc-RD = 198.51.100.1/125, Assoc-SPI = 39, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.7/11], [SI = 253, SFT = 43, RD = 192.0.2.3/11] SFP23: RD = 198.51.100.1/123, SPI = 37, Assoc-Type = 1, Assoc-RD = 198.51.100.1/120, Assoc-SPI = 34, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.5/11], [SI = 253, SFT = 41, RD = 192.0.2.1/11] SFP24: RD = 198.51.100.1/124, SPI = 38, Assoc-Type = 1, Assoc-RD = 198.51.100.1/121, Assoc-SPI = 35, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.6/11], [SI = 253, SFT = 41, RD = 192.0.2.1/11] SFP25: RD = 198.51.100.1/125, SPI = 39, Assoc-Type = 1, Assoc-RD = 198.51.100.1/122, Assoc-SPI = 36, [SI = 255, SFT = 43, RD = 192.0.2.3/11], [SI = 254, SFT = 42, RD = 192.0.2.7/11], [SI = 253, SFT = 41, RD = 192.0.2.1/11]]]> </artwork> </figure>]]></sourcecode> </section> <section anchor="stateegpllchc"title="Parallelnumbered="true" toc="default"> <name>Parallel SFPs Downstream of theChoice" >Choice</name> <t>The mechanism of parallel SFPs demonstrated in <xref target="stateeg2pll"/>format="default"/> is perfectly functional and may be practical in many environments. However, there may be scaling concerns because of the large amount of state (knowledge ofSFPs,SFPs -- i.e., SFPR advertisements retained) if there is a very largeamount of choicenumber of possible SFIs (for example, tens of instances of the same statefulSF),SF) or if there are multiple choices of stateful SF along a path. This situation may be mitigated using SFP fragments that are combined to form theend to endend-to-end SFPs.</t> <t>The example presented here is necessarilysimplistic,simplistic but should convey the basic principle. The example presented in <xref target="eg2pllchcfig"/>format="default"/> is similar to that in <xref target="stateeg2pll"/>format="default"/> but with an additional first hop.</t> <figureanchor="eg2pllchcfig" title="Example Withanchor="eg2pllchcfig"> <name>Example with Parallel SFPs Downstream ofChoice"> <artwork> <![CDATA[Choice</name> <artwork name="" type="" align="left" alt=""><![CDATA[ ------ | SFIa | |SFT=43| ------ ------ ------ | | SFI | | SFI | --------- |SFT=41| |SFT=42| | SFF5 | ------ ------ ..|192.0.2.5|.. | | ..: --------- :.. --------- ---------.: :.--------- ------ | SFF1 | | SFF2 | --------- | SFF3 | -->|Class-|.|192.0.2.1|.|192.0.2.2|....| SFF6 |....|192.0.2.3|--> -->| ifier| --------- ---------: |192.0.2.6| :--------- ------ : --------- : | : | : ------ : ------ : | SFI | :.. | SFIb | ..: |SFT=44| :.. |SFT=43| ..: ------ : ------ : :.---------.: | SFF7 | |192.0.2.7| --------- | ------ | SFIc | |SFT=43| ------]]> </artwork>]]></artwork> </figure> <t>The six SFIs are advertised as follows:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ RD = 192.0.2.1/11, SFT = 41 RD = 192.0.2.2/11, SFT = 42 RD = 192.0.2.5/11, SFT = 43 (for SFIa) RD = 192.0.2.6/11, SFT = 43 (for SFIb) RD = 192.0.2.7/11, SFT = 43 (for SFIc) RD = 192.0.2.3/11, SFT = 44]]> </artwork> </figure>]]></sourcecode> <t>SFF2 is the point at which aload balancingload-balancing choice must be made. So "tail-end" SFPs are constructed as follows. Each takes in a different SFF that provides access to an SF oftypeType 43.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP26: RD = 198.51.100.1/126, SPI = 40, Assoc-Type = 1, Assoc-RD = 198.51.100.1/130, Assoc-SPI = 44, [SI = 255, SFT = 43, RD = 192.0.2.5/11], [SI = 254, SFT = 44, RD = 192.0.2.3/11] SFP27: RD = 198.51.100.1/127, SPI = 41, Assoc-Type = 1, Assoc-RD = 198.51.100.1/131, Assoc-SPI = 45, [SI = 255, SFT = 43, RD = 192.0.2.6/11], [SI = 254, SFT = 44, RD = 192.0.2.3/11] SFP28: RD = 198.51.100.1/128, SPI = 42, Assoc-Type = 1, Assoc-RD = 198.51.100.1/132, Assoc-SPI = 46, [SI = 255, SFT = 43, RD = 192.0.2.7/11], [SI = 254, SFT = 44, RD = 192.0.2.3/11]]]> </artwork> </figure>]]></sourcecode> <t>Now an end-to-end SFP withload balancingload-balancing choice can be constructed as follows. The choice made by SFF2 is expressed in terms of entering one of the three"tail end""tail-end" SFPs.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP29: RD = 198.51.100.1/129, SPI = 43, [SI = 255, SFT = 41, RD = 192.0.2.1/11], [SI = 254, SFT = 42, RD = 192.0.2.2/11], [SI = 253, {SFT = 1, RD = {SPI=40, SI=255, Rsv=0}, RD = {SPI=41, SI=255, Rsv=0}, RD = {SPI=42, SI=255, Rsv=0} } ]]]> </artwork> </figure>]]></sourcecode> <t>Now, despite theload balancingload-balancing choice being madeotherelsewhere than at the initialClassifier,classifier, it is possible for the reverse SFPs to bewell-constructedwell constructed without any ambiguity. The three reverse paths appear as follows.</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFP30: RD = 198.51.100.1/130, SPI = 44, Assoc-Type = 1, Assoc-RD = 198.51.100.1/126, Assoc-SPI = 40, [SI = 255, SFT = 44, RD = 192.0.2.4/11], [SI = 254, SFT = 43, RD = 192.0.2.5/11], [SI = 253, SFT = 42, RD = 192.0.2.2/11], [SI = 252, SFT = 41, RD = 192.0.2.1/11] SFP31: RD = 198.51.100.1/131, SPI = 45, Assoc-Type = 1, Assoc-RD = 198.51.100.1/127, Assoc-SPI = 41, [SI = 255, SFT = 44, RD = 192.0.2.4/11], [SI = 254, SFT = 43, RD = 192.0.2.6/11], [SI = 253, SFT = 42, RD = 192.0.2.2/11], [SI = 252, SFT = 41, RD = 192.0.2.1/11] SFP32: RD = 198.51.100.1/132, SPI = 46, Assoc-Type = 1, Assoc-RD = 198.51.100.1/128, Assoc-SPI = 42, [SI = 255, SFT = 44, RD = 192.0.2.4/11], [SI = 254, SFT = 43, RD = 192.0.2.7/11], [SI = 253, SFT = 42, RD = 192.0.2.2/11], [SI = 252, SFT = 41, RD = 192.0.2.1/11]]]> </artwork> </figure>]]></sourcecode> </section> </section> <section anchor="v6samples"title="Examplesnumbered="true" toc="default"> <name>Examples Using IPv6Addressing">Addressing</name> <t>This section provides several examples using IPv6 addressing. As will be seen from the examples, there is nothing special or clever about using IPv6 addressing rather than IPv4 addressing.</t> <t>The reference network for these IPv6 examples is based on that described at the top of <xref target="example"/>format="default"/> and shown in <xref target="examplefig"/>.</t>format="default"/>.</t> <t>Assume we have a service function overlay network with four SFFs (SFF1, SFF3, SFF3, and SFF4). The SFFs have addresses in the underlay network as follows:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFF1 2001:db8::192:0:2:1 SFF2 2001:db8::192:0:2:2 SFF3 2001:db8::192:0:2:3 SFF4 2001:db8::192:0:2:4]]> </artwork> </figure>]]></sourcecode> <t>Each SFF provides access to some SFIs from the fourService Function Typesservice function types SFT=41, SFT=42, SFT=43, andSFT=44SFT=44, just as before:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ SFF1 SFT=41 and SFT=42 SFF2 SFT=41 and SFT=43 SFF3 SFT=42 and SFT=44 SFF4 SFT=43 and SFT=44]]> </artwork> </figure>]]></sourcecode> <t>The service function network also contains aControllercontroller with address 2001:db8::198:51:100:1.</t> <t>This example service function overlay network is shown in <xref target="eg6fig"/>.</t>format="default"/>.</t> <figureanchor="eg6fig" title="Exampleanchor="eg6fig"> <name>Example Service Function OverlayNetwork"> <artwork> <![CDATA[Network</name> <artwork name="" type="" align="left" alt=""><![CDATA[ ------------------------ | Controller | | 2001:db8::198:51:100:1 | ------------------------ ------ ------ ------ ------ | SFI | | SFI | | SFI | | SFI | |SFT=41| |SFT=42| |SFT=41| |SFT=43| ------ ------ ------ ------ \ / \ / ------------------- ------------------- | SFF1 | | SFF2 | |2001:db8::192:0:2:1| |2001:db8::192:0:2:2| ------------------- ------------------- ---------- Packet --> | | --> Flows --> |Classifier| -->Dest | | --> ---------- ------------------- ------------------- | SFF3 | | SFF4 | |2001:db8::192:0:2:3| |2001:db8::192:0:2:4| ------------------- ------------------- / \ / \ ------ ------ ------ ------ | SFI | | SFI | | SFI | | SFI | |SFT=42| |SFT=44| |SFT=43| |SFT=44| ------ ------ ------ ------]]> </artwork>]]></artwork> </figure> <t>The SFFs advertise routes to the SFIs they support. These advertisements containRoute DistinguishersRDs that are set according to the networkoperator'soperator's configuration model. Note that in an IPv6 network, the RD is not large enough to contain the full IPv6addressaddress, as only six octets areavailable so,available. So, in all of these IPv6 examples, we use RDs oftypeType 1 such that the available six octets are partitioned as four octets for an IPv4 address of the advertising SFF, and two octets that are a local index of the SFI. Furthermore, we have chosen an IPv6 addressing scheme so that thelow orderlow-order four octets of the IPv6 address match an IPv4 address of the advertising node. This scheme is chosen purely for convenience of documentation, and an operator is totally free to use any other scheme so long as it conforms to the definitions of SFIR and SFPR in Sections <xref target="sfiRoutes"/>format="counter"/> and <xref target="sfpRoutes"/>.</t>format="counter"/>.</t> <t>Observant readers will notice that this makes the BGP advertisements shown in these examples exactly the same as in the previous examples. All that is different is that the advertising SFFs andControllercontroller have IPv6 addresses.</t><t>Thus<t>Thus, we see the following SFIRsadvertised:</t>advertised.</t> <t>The SFFs advertise routes to the SFIs they support. So we see the following SFIRs:</t><figure> <artwork> <![CDATA[<sourcecode><![CDATA[ RD = 192.0.2.1/1, SFT = 41 RD = 192.0.2.1/2, SFT = 42 RD = 192.0.2.2/1, SFT = 41 RD = 192.0.2.2/2, SFT = 43 RD = 192.0.2.3/7, SFT = 42 RD = 192.0.2.3/8, SFT = 44 RD = 192.0.2.4/5, SFT = 43 RD = 192.0.2.4/6, SFT = 44]]> </artwork> </figure>]]></sourcecode> <t>Note that the addressing used for communicating between SFFs is taken from theTunnel Encapsulationtunnel encapsulation attribute of the SFIR and not from the SFIR-RD.</t> <section anchor="eg6explicit"title="Examplenumbered="true" toc="default"> <name>Example Explicit SFPWithwith NoChoices" >Choices</name> <t>Consider the following SFPR similar to that in <xref target="exampleexplicit"/>.</t> <figure> <artwork> <![CDATA[format="default"/>.</t> <sourcecode><![CDATA[ SFP1: RD = 198.51.100.1/101, SPI = 15, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, RD = 192.0.2.2/2]]]> </artwork> </figure>]]></sourcecode> <t>TheService Function PathSFP consists of an SF oftypeType 41 located atSFF1SFF1, followed by an SF oftypeType 43 located at SFF2. This path is fullyexplicitexplicit, and each SFF is offered no choice in forwarding a packet along the path.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (15). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 has no flexibility in the choice of SFF to support thenext hopnext-hop SFI and will forward the packet toSFF2SFF2, which will send the packets to the SFI that supports SFT 43 before forwarding the packets to their destinations.</t> </section> <section anchor="eg6choice"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Choice ofSFIs" > <figure> <artwork> <![CDATA[SFIs</name> <sourcecode><![CDATA[ SFP2: RD = 198.51.100.1/102, SPI = 16, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 43, {RD = 192.0.2.2/2, RD = 192.0.2.4/5 } ]]]> </artwork> </figure>]]></sourcecode> <t>In this example, like that in <xref target="examplechoice"/>,format="default"/>, the path also consists of an SF oftypeType 41 located atSFF1SFF1, and this is followed by an SF oftype 43,Type 43; but in thiscasecase, the SI = 250 contains a choice between the SFI located at SFF2 and the SFI located at SFF4.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (16). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a choice ofnext hop SFFnext-hop SFFs to execute the next hop in the path. It can either forward packets to SFF2 or SFF4 to execute a function oftypeType 43. It uses its localload balancingload-balancing algorithm to make this choice. The chosen SFF will send the packets to the SFI that supports SFT 43 before forwarding the packets to their destinations.</t> </section> <section anchor="eg6open"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Open Choice ofSFIs" > <figure> <artwork> <![CDATA[SFIs</name> <sourcecode><![CDATA[ SFP3: RD = 198.51.100.1/103, SPI = 17, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, SFT = 44, RD = 0]]]> </artwork> </figure>]]></sourcecode> <t>In this example, like that in <xref target="exampleopen"/>format="default"/>, the path also consists of an SF oftypeType 41 located atSFF1SFF1, and this is followed by an SI with an RD of zero and SF oftypeType 44. This means that a choice can be made between any SFF that supports an SFI oftypeType 44.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (17). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a free choice ofnext hop SFFnext-hop SFFs to execute the next hop in thepathpath, selecting between all SFFs that support SFs oftypeType 44. Looking at the SFIRs it has received, SFF1 knows that SFtypeType 44 is supported by SFF3 and SFF4. SFF1 uses its localload balancingload-balancing algorithm to make this choice. The chosen SFF will send the packets to the SFI that supports SFT 44 before forwarding the packets to their destinations.</t> </section> <section anchor="eg6sft"title="Examplenumbered="true" toc="default"> <name>Example SFPWithwith Choice ofSFTs" > <figure> <artwork> <![CDATA[SFTs</name> <sourcecode><![CDATA[ SFP4: RD = 198.51.100.1/104, SPI = 18, [SI = 255, SFT = 41, RD = 192.0.2.1/1], [SI = 250, {SFT = 43, RD = 192.0.2.2/2, SFT = 44, RD = 192.0.2.3/8 } ]]]> </artwork> </figure>]]></sourcecode> <t>This example, similar to that in <xref target="examplesft"/>format="default"/>, provides a choice of SF type in the second hop in the path. The SI of 250 indicates a choice between SFtypeType 43 located through SF2 and SFtypeType 44 located at SF3.</t> <t>SFF1 will receive packets on the path from theClassifierclassifier and will identify the path from the SPI (18). The initial SI will be255255, and so SFF1 will deliver the packets to the SFI for SFT 41.</t> <t>When the packets are returned to SFF1 by theSFISFI, the SI will be decreased to 250 for the next hop. SFF1 now has a free choice ofnext hop SFFnext-hop SFFs to execute the next hop in thepathpath, selecting between all SFFs that support an SF oftypeType 43 andSFF3 thatSFF3, which supports an SF oftypeType 44. These may be completely different functions that are to be executed dependent on specific conditions, or they may be similar functions identified with different type identifiers (such as firewalls from different vendors). SFF1 uses its local policy andload balancingload-balancing algorithm to make this choice, and it may use additional information passed back from the local SFI to help inform its selection. The chosen SFF will send the packets to the SFI that supports thechosechosen SFT before forwarding the packets to their destinations.</t> </section> </section> </section> <section anchor="security"title="Security Considerations">numbered="true" toc="default"> <name>Security Considerations</name> <t>The mechanisms in this document use BGP for the control plane. Hence, techniques such as those discussed in <xref target="RFC5925"/>]format="default"/> can be used to help authenticate BGP sessionsand thusand, thus, the messages between BGP peers, making it harder to spoof updates (which could be used to install bogus SFPs ortoadvertise false SIs) or withdrawals.</t> <t>Further discussion of security considerations for BGP may be found in the BGP specification itself <xref target="RFC4271"/>format="default"/> andinthe security analysis for BGP <xref target="RFC4272"/>. The originalformat="default"/>. <xref target="RFC5925" format="default"/> contains a discussion of theuseinappropriateness of the TCP MD5 signature optionto protectfor protecting BGPsessions is found in <xref target="RFC5925" />, whilesessions. <xreftarget="RFC6952" />target="RFC6952"/> includes an analysis of BGP keying and authentication issues.</t> <t>Additionally, this document depends on other documents that specify BGP Multiprotocol Extensions and the documents that define the attributes that are carried by BGP UPDATEs of the SFC AFI/SAFI. <xref target="RFC4760"/>format="default"/> observes that the use of AFI/SAFI does not change the underlying security issues inherent in the existing BGP. Relevant additional security measures are considered in <xreftarget="I-D.ietf-idr-tunnel-encaps" />.</t>target="RFC9012" format="default"/>.</t> <t>This document does not fundamentally change the security behavior of BGP deployments, which depend considerably on the networkoperator'soperator's perception of risk in their network. It may be observed that the application of the mechanisms described in this documentareis scoped to a singledomaindomain, as implied by <xref target="RFC8300"/>format="default"/> and noted in <xref target="funcover"/>format="default"/> of this document. Applicability of BGP within a single domain may enable a network operator to make easier and more consistent decisions about what security measures to apply, and the domain boundary, which BGP enforces by definition, provides a safeguard that prevents leakage of SFC programming in either direction at the boundary.</t> <t>ServiceFunction Chainingfunction chaining provides a significant attackopportunity:opportunity; packets can be diverted from their normal paths through the network, packets can be made to execute unexpected functions, and the functions that are instantiated in software can be subverted. However, this specification does not change the existence ofService Function Chainingservice function chaining, and security issues specific toService Function Chainingservice function chaining are covered in <xref target="RFC7665"/>format="default"/> and <xref target="RFC8300"/>.</t>format="default"/>.</t> <t>This document defines a control plane forService Function Chaining.service function chaining. Clearly, this provides an attack vector for aService Function Chaining systemservice function chaining system, as an attack on this control plane could be used to make the system misbehave. Thus, the security of the BGP system is critically important to the security of the wholeService Function Chainingservice function chaining system. The control plane mechanisms are very similar to those used for BGP/MPLS IP VPNs as described in <xreftarget="RFC4364" />,target="RFC4364"/>, and so the security considerations in that document (Section13)<xref target="RFC4364" section="13" sectionFormat="bare"/>) provide good guidance for securingSFCservice function chaining systems reliant on this specification. Of particular relevance is the need to securely distinguish between messages intended for the control of different SFCoverlaysoverlays, which is similar to the need to distinguish between different VPNs.Section 19 of<xref target="RFC7432"/>section="19" sectionFormat="of"/> also provides useful guidance on the use of BGP in a similar environment.</t> <t>Note that a component ofan SFCa service function chaining system that uses the procedures described in this document also requires communications between aControllercontroller and theSFCservice function chaining network elements (specifically the SFFs andClassifiers).classifiers). This communication covers instructing theClassifiersclassifiers using BGP mechanisms (see <xref target="fspecclassy"/>), thusformat="default"/>); therefore, the use of BGP security is strongly recommended. But it also covers other mechanisms for programming theClassifierclassifier and instructing the SFFs and SFs (for example, to bind SFs to an SFF, and to cause the establishment of tunnels between SFFs). This document does not cover these lattermechanismsmechanisms, and so their security is out of scope, but it should be noted that these communications provide an attack vector on theSFC systemservice function chaining system, and so attention must be paid to ensuring that they are secure.</t> <t>There is an intrinsic assumption inSFCservice function chaining systems that nodes that announce support for specific SFs actually offer thosefunctions,functions and that SFs are not, themselves, attacked or subverted. This is particularly important when the SFs are implemented as software that can be updated. Protection against this sort of concern forms part of the security of anySFCservice function chaining system and so is outside the scope of the control plane mechanisms described in this document.</t> <t>Similarly, there is a vulnerability if a rogue or subvertedControllercontroller announcesSFPsSFPs, especially if that controller "takes over" an existing SFP and changes its contents. Thisiscorresponds to a rogue BGP speaker entering a routing system, or eventoa Route Reflector becoming subverted. Protection mechanisms, as above, include securing BGP sessions and protecting software loads on the controllers.</t> <t>In an environment where there is concern that rogueControllerscontrollers might be introduced to the network and inject false SFPRs or take over and change existing SFPRs, it isRECOMMENDED<bcp14>RECOMMENDED</bcp14> that each SFF andClassifierclassifier be configured with the identities of authorizedControllers.controllers. Thus, the announcement of an SFPR by any other BGP peer would be rejected.</t> <t>Lastly, note that <xref target="sfparules"/>format="default"/> makes two operational suggestions that have implications for the stability and security of the mechanisms described in this document:<list style="symbols"> <t>That</t> <ul spacing="normal"> <li>That modifications to active SFPs not bemade.</t> <t>Thatmade.</li> <li>That SPIs not be immediatelyre-used.</t> </list></t>reused.</li> </ul> </section> <section anchor="iana"title="IANA Considerations">numbered="true" toc="default"> <name>IANA Considerations</name> <section anchor="afisafi"title="Newnumbered="true" toc="default"> <name>New BGPAF/SAFI">AF/SAFI</name> <t>IANA maintainsa registry ofthe "Address FamilyNumbers".Numbers" registry. IANAis requested to assignhas assigned a new Address Family Number from the "Standards Action" range called "BGP SFC"(TBD1 in this document)(31), with this document as a reference.</t> <t>IANA maintainsa registry ofthe "Subsequent Address Family Identifiers (SAFI)Parameters".Parameters" registry. IANAis requested to assignhas assigned a new SAFI value from the "Standards Action" range called "BGP SFC"(TBD2 in this document)(9), with this document as a reference.</t> </section> <section anchor="ianasfpatt"title="Newnumbered="true" toc="default"> <name>"SFP attribute" BGP PathAttribute">Attribute</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP) Parameters" with a subregistry of "BGP Path Attributes". IANAis requested to assignhas assigned a new Path attribute called "SFP attribute"(TBD3 in this document)with a value of 37 and with this document as a reference.</t> </section> <section anchor="ianasftlv"title="New SFPnumbered="true" toc="default"> <name>"SFP AttributeTLVs Type Registry">TLVs" Registry</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP) Parameters". IANAis request to createhas created a new subregistry called the "SFP Attribute TLVs" registry.</t> <t>Valid values are in the range 0 to 65535.<list style="symbols"> <t>Values</t> <ul spacing="normal"> <li>Values 0 and 65535 areto bemarked"Reserved, not to be allocated".</t> <t>Values"Reserved".</li> <li>Values 1 through 65534 are to be assigned according to the "First Come First Served" policy <xref target="RFC8126"/>.</t> </list></t>format="default"/>.</li> </ul> <t>This documentshould be given asis a reference for this registry.</t> <t>Thenewregistryshould track: <list style="symbols"> <t>Type</t> <t>Name</t> <t>Reference Document or Contact</t> <t>Registration Date</t> </list></t>tracks: </t> <ul spacing="normal"> <li>Type</li> <li>Name</li> <li>Reference</li> <li>Registration Date</li> </ul> <t>The registryshouldis initiallybepopulated as follows:</t><figure> <artwork> <![CDATA[ Type | Name | Reference | Date ------+-------------------------+---------------+--------------- 1 | Association TLV | [This.I-D] | Date-to-be-set 2 | Hop TLV | [This.I-D] | Date-to-be-set 3 | SFT TLV | [This.I-D] | Date-to-be-set 4 | MPLS Swapping/Stacking | [This.I-D] | Date-to-be-set 5 | SFP<table> <name>SFP Attribute TLVs Subregistry Initial Contents</name> <thead> <tr> <th>Type</th> <th>Name</th> <th>Reference</th> <th>Registration Date</th> </tr> </thead> <tbody> <tr> <td>1</td> <td>Association TLV</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>2</td> <td>Hop TLV</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>3</td> <td>SFT TLV</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>4</td> <td>MPLS Swapping/Stacking</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>5</td> <td>SFP Traversal WithMPLS | [This.I-D] | Date-to-be-set ]]> </artwork> </figure>MPLS</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> </tbody> </table> </section> <section anchor="ianaassoc"title="New SFPnumbered="true" toc="default"> <name>"SFP AssociationType Registry">Type" Registry</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP) Parameters". IANAis request to createhas created a new subregistry called the "SFP Association Type" registry.</t> <t>Valid values are in the range 0 to 65535.<list style="symbols"> <t>Values</t> <ul spacing="normal"> <li>Values 0 and 65535 areto bemarked"Reserved, not to be allocated".</t> <t>Values"Reserved".</li> <li>Values 1 through 65534 areto beassigned according to the "First Come First Served" policy <xref target="RFC8126"/>.</t> </list></t>format="default"/>.</li> </ul> <t>This documentshould beis given as a reference for this registry.</t> <t>The new registryshould track: <list style="symbols"> <t>Association Type</t> <t>Name</t> <t>Reference Document or Contact</t> <t>Registration Date</t> </list></t>tracks: </t> <ul spacing="normal"> <li>Association Type</li> <li>Name</li> <li>Reference</li> <li>Registration Date</li> </ul> <t>The registry should initially be populated as follows:</t><figure> <artwork> <![CDATA[<table> <name>SFP Association Type| Name | Reference | Date -----------------+--------------------+------------+--------------- 1 | Bidirectional SFP | [This.I-D] | Date-to-be-set ]]> </artwork> </figure>Subregistry Initial Contents</name> <thead> <tr> <th>Association Type</th> <th>Name</th> <th>Reference</th> <th>Date</th> </tr> </thead> <tbody> <tr> <td>1</td> <td>Bidirectional SFP</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> </tbody> </table> </section> <section anchor="SFTreg"title="Newnumbered="true" toc="default"> <name>"Service Function Chaining Service FunctionType Registry">Types" Registry</name> <t>IANAis request to createhas created a new top-level registry called "Service Function Chaining Service Function Types".</t> <t>Valid values are in the range 0 to 65535.<list style="symbols"> <t>Values</t> <ul spacing="normal"> <li>Values 0 and 65535 areto bemarked"Reserved, not to be allocated".</t> <t>Values"Reserved".</li> <li>Values 1 through 31 are to be assigned by "Standards Action" <xref target="RFC8126"/>format="default"/> and are referred to as theSpecial Purpose"special-purpose SFTvalues.</t> <t>Valuesvalues".</li> <li>Values 32 through 64495 are to be assigned according to the "First Come First Served" policy <xref target="RFC8126"/>.</t> <t>Valuesformat="default"/>.</li> <li>Values 64496 through 65534 are for Private Use and are not to be recorded byIANA.</t> </list></t>IANA.</li> </ul> <t>This documentshould beis given as a reference for this registry.</t> <t>Thenewregistryshould track: <list style="symbols"> <t>Value</t> <t>Name</t> <t>Reference Document or Contact</t> <t>Registration Date</t> </list></t>tracks: </t> <ul spacing="normal"> <li>Value</li> <li>Name</li> <li>Reference</li> <li>Registration Date</li> </ul> <t>The registryshouldis initiallybepopulated asfollows where [I-D.darwa] should be expanded to <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments" />.</t> <figure> <artwork> <![CDATA[ Value | Name | Reference | Date ------+-------------------------+------------+--------------- 0 | Reserved, not to be | [This.I-D] | Date-to-be-set | allocated | | 1 | Change Sequence | [This.I-D] | Date-to-be-set 2-31 | Unassigned | | 32 | Classifier | [This.I-D] | Date-to-be-set | | [I-D.dawra]| 33 | Firewall | [This.I-D] | Date-to-be-set | | [I-D.dawra]| 34 | Load balancer | [This.I-D] | Date-to-be-set | | [I-D.dawra]| 35 | Deepfollows.</t> <table> <name>Service Function Chaining Service Function Types Registry Initial Contents</name> <thead> <tr> <th>Value</th> <th>Name</th> <th>Reference</th> <th>Date</th> </tr> </thead> <tbody> <tr> <td>0</td> <td>Reserved</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>1</td> <td>Change Sequence</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>2-31</td> <td>Unassigned</td> <td></td> <td></td> </tr> <tr> <td>32</td> <td>Classifier</td> <td>RFC 9015, <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments" format="default"/></td> <td>2020-09-02</td> </tr> <tr> <td>33</td> <td>Firewall</td> <td>RFC 9015, <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments" format="default"/></td> <td>2020-09-02</td> </tr> <tr> <td>34</td> <td>Load balancer</td> <td>RFC 9015, <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments" format="default"/></td> <td>2020-09-02</td> </tr> <tr> <td>35</td> <td>Deep packet inspection| [This.I-D] | Date-to-be-set | engine | [I-D.dawra]| 36 | Penalty box | [This.I-D] | Date-to-be-set | | [RFC8300] | 37 | WAN accelerator | [This.I-D] | Date-to-be-set | | [RFC7665] | | | [RFC8300] | 38 | Application accelerator | [This.I-D] | Date-to-be-set | | [RFC7665] | 39 | TCP optimizer | [This.I-D] | Date-to-be-set | | [RFC7665] | 40 | Networkengine</td> <td>RFC 9015, <xref target="I-D.dawra-idr-bgp-ls-sr-service-segments" format="default"/></td> <td>2020-09-02</td> </tr> <tr> <td>36</td> <td>Penalty box</td> <td>RFC 9015, <xref target="RFC8300"/></td> <td>2020-09-02</td> </tr> <tr> <td>37</td> <td>WAN accelerator</td> <td>RFC 9015, <xref target="RFC7665"/>, <xref target="RFC8300"/></td> <td>2020-09-02</td> </tr> <tr> <td>38</td> <td>Application accelerator</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>39</td> <td>TCP optimizer</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>40</td> <td>Network Address| [This.I-D] | Date-to-be-set | Translator | [RFC7665] | 41 | NAT44 | [This.I-D] | Date-to-be-set | | [RFC7665] | | | [RFC3022] | 42 | NAT64 | [This.I-D] | Date-to-be-set | | [RFC7665] | | | [RFC6146] | 43 | NPTv6 | [This.I-D] | Date-to-be-set | | [RFC7665] | | | [RFC6296] | 44 | Lawful intercept | [This.I-D] | Date-to-be-set | | [RFC7665] | 45 | HOST_ID injection | [This.I-D] | Date-to-be-set | | [RFC7665] | 46 | HTTPTranslator</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>41</td> <td>NAT44</td> <td>RFC 9015, <xref target="RFC7665"/>, <xref target="RFC3022"/></td> <td>2020-09-02</td> </tr> <tr> <td>42</td> <td>NAT64</td> <td>RFC 9015, <xref target="RFC7665"/>, <xref target="RFC6146"/></td> <td>2020-09-02</td> </tr> <tr> <td>43</td> <td>NPTv6</td> <td>RFC 9015, <xref target="RFC7665"/>, <xref target="RFC6296"/></td> <td>2020-09-02</td> </tr> <tr> <td>44</td> <td>Lawful intercept</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>45</td> <td>HOST_ID injection</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>46</td> <td>HTTP headerenrichment | [This.I-D] | Date-to-be-set | | [RFC7665] | 47 | Cachingenrichment</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td> </tr> <tr> <td>47</td> <td>Caching engine| [This.I-D] | Date-to-be-set | | [RFC7665] | 48- | | | -65534|Unassigned | | 65535 | Reserved,</td> <td>RFC 9015, <xref target="RFC7665"/></td> <td>2020-09-02</td></tr> <tr> <td>48-64495</td> <td>Unassigned</td> <td></td> <td></td> </tr> <tr> <td>64496-65534</td> <td>Reserved for Private Use</td> <td></td> <td></td> </tr> <tr> <td>65535</td> <td>Reserved, not to be| | |allocated| [This.I-D] | Date-to-be-set ]]> </artwork> </figure></td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> </tbody> </table> </section> <section anchor="ExpExtComreg"title="New Generic Transitive Experimental Use Extended Community Sub-Types">numbered="true" toc="default"> <name>Flow Specification for SFC Classifiers</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP)Parameters"Extended Communities" with a subregistry of "Generic Transitive Experimental Use Extended CommunitySub-Type".Sub-Types". IANAis requested to assignhas assigned a newsub-typesubtype as follows:<list style="symbls"> <t>"Flow</t> <ul empty="true" spacing="normal"> <li>"Flow Specification for SFC Classifiers"(TBD4 in this document)with a value of 0x0d and with this document as thereference.</t> </list></t>reference.</li> </ul> </section> <section anchor="TransExtComreg"title="Newnumbered="true" toc="default"> <name>New BGP Transitive Extended CommunityType">Type</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP)Parameters"Extended Communities" with a subregistry of "BGP Transitive Extended Community Types". IANAis requested to assignhas assigned a new type as follows:<list style="symbols"> <t>SFC</t> <ul empty="true" spacing="normal"> <li>SFC (Sub-Types are defined in the "SFC Extended Community Sub-Types" registry)(TBD6 in this document)with a value of 0x0b and with this document as thereference.</t> </list></t>reference.</li> </ul> </section> <section anchor="SFCExtComreg"title="New SFCnumbered="true" toc="default"> <name>"SFC Extended CommunitySub-Types Registry">Sub-Types" Registry</name> <t>IANA maintains a registry of "Border Gateway Protocol (BGP) Parameters". IANAis requested to createhas created a newsub-registrysubregistry called the "SFC Extended CommunitySub-Types Registry".</t>Sub-Types" registry.</t> <t>IANAshould includehas included the followingnote replacing the string "TBD6" with the value assigned for <xref target="TransExtComreg" />: <list style="none"> <t>Thisnote: </t> <aside><t> This registry contains values of the second octet (the "Sub-Type" field) of an extended community when the value of the first octet (the "Type" field) is set toTBD6.</t> </list></t>0x0b.</t> </aside> <t>The allocation policy for this registryshould beis First Come First Served.</t> <t>Valid values are 0 to 255. The value 0 is reserved and should not be allocated.</t> <t>IANAis requested to populatehas populated this registry with the following entries:</t><figure> <artwork> <![CDATA[ Sub-Type | | | Value | Name | Reference | Date ---------+----------------------+-------------+--------------- 0 | Reserved, not to be | | | allocated | | 1 | SFIR Pool Identifier | [This.I-D] | Date-to-be-set 2 | MPLS<table> <name>SFC Extended Community Sub-Types Subregistry Initial Contents</name> <thead> <tr> <th>Sub-Type Value</th> <th>Name</th> <th>Reference</th> <th>Date</th> </tr> </thead> <tbody> <tr> <td>0</td> <td>Reserved</td> <td>RFC 9015</td> <td></td> </tr> <tr> <td>1</td> <td>SFIR pool identifier</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>2</td> <td>MPLS Label Stack| [This.I-D] | Date-to-be-set |MixedSwapping/ | | | Stacking Labels | | 3-255 | Unassigned | | ]]> </artwork> </figure> <t>All other values should be marked "Unassigned".</t>Swapping/Stacking Labels</td> <td>RFC 9015</td> <td>2020-09-02</td> </tr> <tr> <td>3-255</td> <td>Unassigned</td> <td></td> <td></td> </tr> </tbody> </table> </section> <section anchor="SpiSiRep"title="SPI/SI Representation">numbered="true" toc="default"> <name>New SPI/SI Representation Sub-TLV</name> <t>IANAis requested to assignhas assigned a codepoint from the "BGP Tunnel Encapsulation Attribute Sub-TLVs" registry for the "SPI/SI Representation Sub-TLV"(TBD5 in this document)with a value of 16 and with this documentbeingas the reference.</t> </section> <section anchor="IANAbits"title="SFCnumbered="true" toc="default"> <name>"SFC SPI/SI RepresentationFlags Registry">Flags" Registry</name> <t>IANA maintains the "BGP Tunnel Encapsulation Attribute Sub-TLVs" registry andis requested to createhas created an associated registry called the "SFC SPI/SI Representation Flags" registry.</t> <t>Bits are to be assigned by Standards Action. The field is 16 bits long, and bits are counted from thethemost significant bit as bit zero.</t> <t>IANAis requested to populatehas populated the registry as follows:</t><figure> <artwork> <![CDATA[ Bit number | Name | Reference -----------+----------------------+----------- TBD9 | NSH<table> <name>SFC SPI/SI Representation Flags Registry Initial Contents</name> <thead> <tr> <th>Value</th> <th>Name</th> <th>Reference</th> </tr> </thead> <tbody> <tr> <td>0</td> <td>NSH dataplane | [This.I-D] TBD10 | MPLSplane</td> <td>RFC 9015</td> </tr> <tr> <td>1</td> <td>MPLS dataplane | [This.I-D] ]]> </artwork> </figure> </section>plane</td> <td>RFC 9015</td> </tr> </tbody> </table> </section><section anchor="contributors" title="Contributors"> <figure> <artwork> <![CDATA[ Stuart Mackie Juniper Networks Email: wsmackie@juinper.net Keyur Patel Arrcus, Inc. Email: keyur@arrcus.com Avinash Lingala AT&T Email: ar977m@att.com ]]> </artwork> </figure></section> </middle> <back> <displayreference target="I-D.dawra-idr-bgp-ls-sr-service-segments" to="BGP-LS-SR"/> <references> <name>References</name> <references> <name>Normative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2119.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4271.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4360.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4364.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4760.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7432.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7606.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7665.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8126.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8174.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8300.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8595.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8596.xml"/> <!--[I-D.ietf-idr-rfc5575bis]; Pub'd as RFC 8955--> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8955.xml"/> <!--[I-D.ietf-idr-tunnel-encaps]; Pub'd as RFC 9012--> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.9012.xml"/> </references> <references> <name>Informative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3022.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4272.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5925.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6146.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6296.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6952.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7498.xml"/> <!-- [I-D.dawra-idr-bgp-ls-sr-service-segments] IESG state I-D Exists --> <xi:include href="https://datatracker.ietf.org/doc/bibxml3/reference.I-D.dawra-idr-bgp-ls-sr-service-segments.xml"/> </references> </references> <section anchor="acks"title="Acknowledgements">numbered="false" toc="default"> <name>Acknowledgements</name> <t>Thanks toTony Przygienda, Jeff Haas, and Andy Malis<contact fullname="Tony Przygienda"/>, <contact fullname="Jeff Haas"/>, and <contact fullname="Andy Malis"/> for helpful comments, and toJoel Halpern<contact fullname="Joel Halpern"/> for discussions that improved this document.Yuanlong Jiang<contact fullname="Yuanlong Jiang"/> provided a useful review and caught some important issues.Stephane Litkowski<contact fullname="Stephane Litkowski"/> did an exceptionally good and detaileddocument shepherdDocument Shepherd review.</t><t>Andy Malis<t><contact fullname="Andy Malis"/> contributed text that formed the basis of <xref target="mpls-encaps"/>.</t> <t>Brian Carpenter and Martin Vigoureuxformat="default"/>.</t> <t><contact fullname="Brian Carpenter"/> and <contact fullname="Martin Vigoureux"/> provided useful reviews during IETFlast call.Last Call. Thanks also toSheng Jiang, Med Boucadair, Ravi Singh, Benjamin Kaduk, Roman Danyliw, Adam Roach, Alvaro Retana, Barry Leiba, and Murray Kucherawy<contact fullname="Sheng Jiang"/>, <contact fullname="Med Boucadair"/>, <contact fullname="Ravi Singh"/>, <contact fullname="Benjamin Kaduk"/>, <contact fullname="Roman Danyliw"/>, <contact fullname="Adam Roach"/>, <contact fullname="Alvaro Retana"/>, <contact fullname="Barry Leiba"/>, and <contact fullname="Murray Kucherawy"/> for review comments.Ketan Talaulikar<contact fullname="Ketan Talaulikar"/> provided helpful discussion of the SFTcode pointcodepoint registry.Ron Bonica<contact fullname="Ron Bonica"/> kept us honest on the difference between an RD and an RT;Benjamin Kaduk<contact fullname="Benjamin Kaduk"/> kept us on message about thedifferncedifference between an RD and anextended community.</t>Extended Community.</t> </section> <section anchor="contributors" numbered="false" toc="default"> <name>Contributors</name> <contact fullname="Stuart Mackie"> <organization>Juniper Networks</organization> <address> <email>wsmackie@juinper.net</email> </address> </contact> <contact fullname="Keyur Patel"> <organization>Arrcus, Inc.</organization> <address> <email>keyur@arrcus.com</email> </address> </contact> <contact fullname="Avinash Lingala"> <organization>AT&T</organization> <address> <email>ar977m@att.com</email> </address> </contact> </section></middle> <back> <references title="Normative References"> &RFC2119; &RFC4271; &RFC4360; &RFC4364; &RFC4760; &RFC7432; &RFC7606; &RFC7665; &RFC8126; &RFC8174; &RFC8300; &RFC8595; &RFC8596; <?rfc include="reference.I-D.ietf-idr-rfc5575bis"?> <?rfc include="reference.I-D.ietf-idr-tunnel-encaps"?> </references> <references title="Informative References"> &RFC3022; &RFC4272; &RFC5925; &RFC6146; &RFC6296; &RFC6952; &RFC7498; <?rfc include="reference.I-D.dawra-idr-bgp-ls-sr-service-segments"?> </references></back> </rfc>