rfc8662xml2.original.xml   rfc8662.xml 
<?xml version="1.0" encoding="US-ASCII"?> <?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc SYSTEM "rfc2629.dtd"[ <rfc xmlns:xi="http://www.w3.org/2001/XInclude" version="3" category="std" conse
<!ENTITY RFC2119 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen nsus="true" docName="draft-ietf-mpls-spring-entropy-label-12" indexInclude="true
ce.RFC.2119.xml"> " ipr="trust200902" number="8662" prepTime="2019-12-04T22:14:22" scripts="Common
<!ENTITY RFC6790 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen ,Latin" sortRefs="false" submissionType="IETF" symRefs="true" tocDepth="3" tocIn
ce.RFC.6790.xml"> clude="true" xml:lang="en">
<!ENTITY RFC4206 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen <link href="https://datatracker.ietf.org/doc/draft-ietf-mpls-spring-entropy-la
ce.RFC.4206.xml"> bel-12" rel="prev"/>
<!ENTITY RFC7325 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen <link href="https://dx.doi.org/10.17487/rfc8662" rel="alternate"/>
ce.RFC.7325.xml"> <link href="urn:issn:2070-1721" rel="alternate"/>
<!ENTITY RFC7855 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen
ce.RFC.7855.xml">
<!ENTITY RFC8174 SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml/referen
ce.RFC.8174.xml">
<!ENTITY SR SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I
-D.ietf-spring-segment-routing.xml">
<!ENTITY SR-MPLS SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refere
nce.I-D.ietf-spring-segment-routing-mpls.xml">
<!ENTITY ISIS-ELC SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refer
ence.I-D.ietf-isis-mpls-elc.xml">
<!ENTITY OSPF-ELC SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refer
ence.I-D.ietf-ospf-mpls-elc.xml">
<!ENTITY OSPF-MSD SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refer
ence.I-D.ietf-ospf-segment-routing-msd.xml">
<!ENTITY ISIS-MSD SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refer
ence.I-D.ietf-isis-segment-routing-msd.xml">
<!ENTITY BGP-MSD SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/refere
nce.I-D.ietf-idr-bgp-ls-segment-routing-msd.xml">
<!ENTITY SR-L2-BUNDLES SYSTEM "http://xml2rfc.tools.ietf.org/public/rfc/bibxml3/
reference.I-D.ietf-isis-l2bundles.xml">
]>
<?rfc toc="yes" ?>
<?rfc tocompact="yes"?>
<?rfc tocdepth="4"?>
<?rfc tocindent="yes"?>
<?rfc symrefs="yes" ?>
<?rfc sortrefs="no"?>
<?rfc rfcedstyle="yes"?>
<?rfc subcompact="no"?>
<?rfc compact="yes" ?>
<?rfc iprnotified="Yes" ?>
<?rfc strict="no" ?>
<rfc ipr="trust200902" category="std" docName="draft-ietf-mpls-spring-entropy-la
bel-12" obsoletes="" updates="" submissionType="IETF" xml:lang="en">
<front> <front>
<title abbrev="Entropy Labels for SPRING tunnels">Entropy label for SPRING t <title abbrev="Entropy Labels for SPRING Tunnels">Entropy Label for Source P
unnels</title> acket Routing in Networking (SPRING) Tunnels</title>
<seriesInfo name="RFC" value="8662" stream="IETF"/>
<author initials="S" surname="Kini" fullname="Sriganesh Kini"> <author initials="S" surname="Kini" fullname="Sriganesh Kini">
<organization></organization> <organization showOnFrontPage="true"/>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>sriganeshkini@gmail.com</email> <email>sriganeshkini@gmail.com</email>
</address> </address>
</author> </author>
<author initials="K" surname="Kompella" fullname="Kireeti Kompella"> <author initials="K" surname="Kompella" fullname="Kireeti Kompella">
<organization>Juniper</organization> <organization showOnFrontPage="true">Juniper</organization>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>kireeti@juniper.net</email> <email>kireeti@juniper.net</email>
</address> </address>
</author> </author>
<author initials="S" surname="Sivabalan" fullname="Siva Sivabala <author initials="S" surname="Sivabalan" fullname="Siva Sivabalan">
n"> <organization showOnFrontPage="true">Cisco</organization>
<organization>Cisco</organization>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>msiva@cisco.com</email> <email>msiva@cisco.com</email>
</address> </address>
</author> </author>
<author initials="S" surname="Litkowski" fullname="Stephane Litk <author initials="S" surname="Litkowski" fullname="Stephane Litkowski">
owski"> <organization showOnFrontPage="true">Orange</organization>
<organization>Orange</organization>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>stephane.litkowski@orange.com</email> <email>slitkows.ietf@gmail.com</email>
</address> </address>
</author> </author>
<author initials="R" surname="Shakir" fullname="Rob Shakir"> <author initials="R" surname="Shakir" fullname="Rob Shakir">
<organization>Google</organization> <organization showOnFrontPage="true">Google</organization>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>rjs@rob.sh</email> <email>robjs@google.com</email>
</address> </address>
</author> </author>
<author initials="J" surname="Tantsura" fullname="Jeff Tantsura" <author initials="J" surname="Tantsura" fullname="Jeff Tantsura">
> <organization showOnFrontPage="true">Apstra, Inc.</organization>
<organization></organization>
<address> <address>
<postal> <postal>
<street></street> <street/>
<city></city> <city/>
<region></region> <region/>
<code></code> <code/>
<country></country> <country/>
</postal> </postal>
<email>jefftant@gmail.com</email> <email>jefftant.ietf@gmail.com</email>
</address> </address>
</author> </author>
<date month="12" year="2019"/>
<date year="2018" />
<area>Routing</area> <area>Routing</area>
<workgroup>Network Working Group</workgroup> <keyword>Flow-aware load balancing</keyword>
<abstract> <keyword>ECMP</keyword>
<keyword>equal-cost multipath</keyword>
<t> <abstract pn="section-abstract">
Segment Routing (SR) leverages the source routing paradigm. A node <t pn="section-abstract-1">
steers a packet through an ordered list of instructions, called Segment Routing (SR) leverages the source-routing paradigm. A node steers a
segments. Segment Routing can be applied to the Multi Protocol Label packet through an ordered list of instructions, called segments. Segment
Switching (MPLS) data plane. Entropy label (EL) is a technique used Routing can be applied to the Multiprotocol Label Switching (MPLS) data
in MPLS to improve load-balancing. This document examines and plane. Entropy labels (ELs) are used in MPLS to improve load-balancing.
describes how ELs are to be applied to Segment Routing MPLS. This document examines and describes how ELs are to be applied to Segment
Routing MPLS.
</t> </t>
</abstract> </abstract>
<boilerplate>
<section anchor="status-of-memo" numbered="false" removeInRFC="false" toc=
"exclude" pn="section-boilerplate.1">
<name slugifiedName="name-status-of-this-memo">Status of This Memo</name
>
<t pn="section-boilerplate.1-1">
This is an Internet Standards Track document.
</t>
<t pn="section-boilerplate.1-2">
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by
the Internet Engineering Steering Group (IESG). Further
information on Internet Standards is available in Section 2 of
RFC 7841.
</t>
<t pn="section-boilerplate.1-3">
Information about the current status of this document, any
errata, and how to provide feedback on it may be obtained at
<eref target="https://www.rfc-editor.org/info/rfc8662" brackets="non
e"/>.
</t>
</section>
<section anchor="copyright" numbered="false" removeInRFC="false" toc="excl
ude" pn="section-boilerplate.2">
<name slugifiedName="name-copyright-notice">Copyright Notice</name>
<t pn="section-boilerplate.2-1">
Copyright (c) 2019 IETF Trust and the persons identified as the
document authors. All rights reserved.
</t>
<t pn="section-boilerplate.2-2">
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(<eref target="https://trustee.ietf.org/license-info" brackets="none
"/>) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with
respect to this document. Code Components extracted from this
document must include Simplified BSD License text as described in
Section 4.e of the Trust Legal Provisions and are provided without
warranty as described in the Simplified BSD License.
</t>
</section>
</boilerplate>
<toc>
<section anchor="toc" numbered="false" removeInRFC="false" toc="exclude" p
n="section-toc.1">
<name slugifiedName="name-table-of-contents">Table of Contents</name>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="section-to
c.1-1">
<li pn="section-toc.1-1.1">
<t keepWithNext="true" pn="section-toc.1-1.1.1"><xref derivedContent
="1" format="counter" sectionFormat="of" target="section-1"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-introduction">Introductio
n</xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="sectio
n-toc.1-1.1.2">
<li pn="section-toc.1-1.1.2.1">
<t keepWithNext="true" pn="section-toc.1-1.1.2.1.1"><xref derive
dContent="1.1" format="counter" sectionFormat="of" target="section-1.1"/>.  <xre
f derivedContent="" format="title" sectionFormat="of" target="name-requirements-
language">Requirements Language</xref></t>
</li>
</ul>
</li>
<li pn="section-toc.1-1.2">
<t keepWithNext="true" pn="section-toc.1-1.2.1"><xref derivedContent
="2" format="counter" sectionFormat="of" target="section-2"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-abbreviations-and-termino
lo">Abbreviations and Terminology</xref></t>
</li>
<li pn="section-toc.1-1.3">
<t keepWithNext="true" pn="section-toc.1-1.3.1"><xref derivedContent
="3" format="counter" sectionFormat="of" target="section-3"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-use-case-requiring-multip
at">Use Case Requiring Multipath Load-Balancing</xref></t>
</li>
<li pn="section-toc.1-1.4">
<t keepWithNext="true" pn="section-toc.1-1.4.1"><xref derivedContent
="4" format="counter" sectionFormat="of" target="section-4"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-entropy-readable-label-de
pt">Entropy Readable Label Depth</xref></t>
</li>
<li pn="section-toc.1-1.5">
<t keepWithNext="true" pn="section-toc.1-1.5.1"><xref derivedContent
="5" format="counter" sectionFormat="of" target="section-5"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-maximum-sid-depth">Maximu
m SID Depth</xref></t>
</li>
<li pn="section-toc.1-1.6">
<t keepWithNext="true" pn="section-toc.1-1.6.1"><xref derivedContent
="6" format="counter" sectionFormat="of" target="section-6"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-lsp-stitching-using-the-b
in">LSP Stitching Using the Binding SID</xref></t>
</li>
<li pn="section-toc.1-1.7">
<t keepWithNext="true" pn="section-toc.1-1.7.1"><xref derivedContent
="7" format="counter" sectionFormat="of" target="section-7"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-insertion-of-entropy-labe
ls">Insertion of Entropy Labels for SPRING Path</xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="sectio
n-toc.1-1.7.2">
<li pn="section-toc.1-1.7.2.1">
<t keepWithNext="true" pn="section-toc.1-1.7.2.1.1"><xref derive
dContent="7.1" format="counter" sectionFormat="of" target="section-7.1"/>.  <xre
f derivedContent="" format="title" sectionFormat="of" target="name-overview">Ove
rview</xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="se
ction-toc.1-1.7.2.1.2">
<li pn="section-toc.1-1.7.2.1.2.1">
<t keepWithNext="true" pn="section-toc.1-1.7.2.1.2.1.1"><xre
f derivedContent="7.1.1" format="counter" sectionFormat="of" target="section-7.1
.1"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-e
xample-1-the-ingress-node-">Example 1: The Ingress Node Has a Sufficient MSD</xr
ef></t>
</li>
<li pn="section-toc.1-1.7.2.1.2.2">
<t keepWithNext="true" pn="section-toc.1-1.7.2.1.2.2.1"><xre
f derivedContent="7.1.2" format="counter" sectionFormat="of" target="section-7.1
.2"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-e
xample-2-the-ingress-node-">Example 2: The Ingress Node Does Not Have a Sufficie
nt MSD</xref></t>
</li>
</ul>
</li>
<li pn="section-toc.1-1.7.2.2">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.1"><xref derive
dContent="7.2" format="counter" sectionFormat="of" target="section-7.2"/>.  <xre
f derivedContent="" format="title" sectionFormat="of" target="name-consideration
s-for-the-plac">Considerations for the Placement of Entropy Labels</xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="se
ction-toc.1-1.7.2.2.2">
<li pn="section-toc.1-1.7.2.2.2.1">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.2.1.1"><xre
f derivedContent="7.2.1" format="counter" sectionFormat="of" target="section-7.2
.1"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-e
rld-value">ERLD Value</xref></t>
</li>
<li pn="section-toc.1-1.7.2.2.2.2">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.2.2.1"><xre
f derivedContent="7.2.2" format="counter" sectionFormat="of" target="section-7.2
.2"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-s
egment-type">Segment Type</xref></t>
</li>
<li pn="section-toc.1-1.7.2.2.2.3">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.2.3.1"><xre
f derivedContent="7.2.3" format="counter" sectionFormat="of" target="section-7.2
.3"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-m
aximizing-number-of-lsrs-t">Maximizing Number of LSRs That Will Load-Balance</xr
ef></t>
</li>
<li pn="section-toc.1-1.7.2.2.2.4">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.2.4.1"><xre
f derivedContent="7.2.4" format="counter" sectionFormat="of" target="section-7.2
.4"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-p
reference-for-a-part-of-th">Preference for a Part of the Path</xref></t>
</li>
<li pn="section-toc.1-1.7.2.2.2.5">
<t keepWithNext="true" pn="section-toc.1-1.7.2.2.2.5.1"><xre
f derivedContent="7.2.5" format="counter" sectionFormat="of" target="section-7.2
.5"/>.  <xref derivedContent="" format="title" sectionFormat="of" target="name-c
ombining-criteria">Combining Criteria</xref></t>
</li>
</ul>
</li>
</ul>
</li>
<li pn="section-toc.1-1.8">
<t keepWithNext="true" pn="section-toc.1-1.8.1"><xref derivedContent
="8" format="counter" sectionFormat="of" target="section-8"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-a-simple-example-algorith
m">A Simple Example Algorithm</xref></t>
</li>
<li pn="section-toc.1-1.9">
<t keepWithNext="true" pn="section-toc.1-1.9.1"><xref derivedContent
="9" format="counter" sectionFormat="of" target="section-9"/>.  <xref derivedCon
tent="" format="title" sectionFormat="of" target="name-deployment-considerations
">Deployment Considerations</xref></t>
</li>
<li pn="section-toc.1-1.10">
<t keepWithNext="true" pn="section-toc.1-1.10.1"><xref derivedConten
t="10" format="counter" sectionFormat="of" target="section-10"/>. <xref derivedC
ontent="" format="title" sectionFormat="of" target="name-options-considered">Opt
ions Considered</xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="sectio
n-toc.1-1.10.2">
<li pn="section-toc.1-1.10.2.1">
<t keepWithNext="true" pn="section-toc.1-1.10.2.1.1"><xref deriv
edContent="10.1" format="counter" sectionFormat="of" target="section-10.1"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-single-el-
at-the-bottom-of-">Single EL at the Bottom of the Stack</xref></t>
</li>
<li pn="section-toc.1-1.10.2.2">
<t keepWithNext="true" pn="section-toc.1-1.10.2.2.1"><xref deriv
edContent="10.2" format="counter" sectionFormat="of" target="section-10.2"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-an-el-per-
segment-in-the-st">An EL per Segment in the Stack</xref></t>
</li>
<li pn="section-toc.1-1.10.2.3">
<t keepWithNext="true" pn="section-toc.1-1.10.2.3.1"><xref deriv
edContent="10.3" format="counter" sectionFormat="of" target="section-10.3"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-a-reusable
-el-for-a-stack-o">A Reusable EL for a Stack of Tunnels</xref></t>
</li>
<li pn="section-toc.1-1.10.2.4">
<t keepWithNext="true" pn="section-toc.1-1.10.2.4.1"><xref deriv
edContent="10.4" format="counter" sectionFormat="of" target="section-10.4"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-el-at-top-
of-stack">EL at Top of Stack</xref></t>
</li>
<li pn="section-toc.1-1.10.2.5">
<t keepWithNext="true" pn="section-toc.1-1.10.2.5.1"><xref deriv
edContent="10.5" format="counter" sectionFormat="of" target="section-10.5"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-els-at-rea
dable-label-stack">ELs at Readable Label Stack Depths</xref></t>
</li>
</ul>
</li>
<li pn="section-toc.1-1.11">
<t keepWithNext="true" pn="section-toc.1-1.11.1"><xref derivedConten
t="11" format="counter" sectionFormat="of" target="section-11"/>. <xref derivedC
ontent="" format="title" sectionFormat="of" target="name-iana-considerations">IA
NA Considerations</xref></t>
</li>
<li pn="section-toc.1-1.12">
<t keepWithNext="true" pn="section-toc.1-1.12.1"><xref derivedConten
t="12" format="counter" sectionFormat="of" target="section-12"/>. <xref derivedC
ontent="" format="title" sectionFormat="of" target="name-security-considerations
">Security Considerations</xref></t>
</li>
<li pn="section-toc.1-1.13">
<t keepWithNext="true" pn="section-toc.1-1.13.1"><xref derivedConten
t="13" format="counter" sectionFormat="of" target="section-13"/>. <xref derivedC
ontent="" format="title" sectionFormat="of" target="name-references">References<
/xref></t>
<ul bare="true" empty="true" indent="2" spacing="compact" pn="sectio
n-toc.1-1.13.2">
<li pn="section-toc.1-1.13.2.1">
<t keepWithNext="true" pn="section-toc.1-1.13.2.1.1"><xref deriv
edContent="13.1" format="counter" sectionFormat="of" target="section-13.1"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-normative-
references">Normative References</xref></t>
</li>
<li pn="section-toc.1-1.13.2.2">
<t keepWithNext="true" pn="section-toc.1-1.13.2.2.1"><xref deriv
edContent="13.2" format="counter" sectionFormat="of" target="section-13.2"/>.  <
xref derivedContent="" format="title" sectionFormat="of" target="name-informativ
e-references">Informative References</xref></t>
</li>
</ul>
</li>
<li pn="section-toc.1-1.14">
<t keepWithNext="true" pn="section-toc.1-1.14.1"><xref derivedConten
t="" format="none" sectionFormat="of" target="section-appendix.a"/><xref derived
Content="" format="title" sectionFormat="of" target="name-acknowledgements">Ackn
owledgements</xref></t>
</li>
<li pn="section-toc.1-1.15">
<t keepWithNext="true" pn="section-toc.1-1.15.1"><xref derivedConten
t="" format="none" sectionFormat="of" target="section-appendix.b"/><xref derived
Content="" format="title" sectionFormat="of" target="name-contributors">Contribu
tors</xref></t>
</li>
<li pn="section-toc.1-1.16">
<t keepWithNext="true" pn="section-toc.1-1.16.1"><xref derivedConten
t="" format="none" sectionFormat="of" target="section-appendix.c"/><xref derived
Content="" format="title" sectionFormat="of" target="name-authors-addresses">Aut
hors' Addresses</xref></t>
</li>
</ul>
</section>
</toc>
</front> </front>
<middle> <middle>
<section title="Introduction" toc="default"> <section toc="include" numbered="true" removeInRFC="false" pn="section-1">
<t> <name slugifiedName="name-introduction">Introduction</name>
Segment Routing <xref target="I-D.ietf-spring-segment-routing"/> is based on <t pn="section-1-1">
source routed tunnels Segment Routing <xref target="RFC8402" format="default" sectionFormat="of" de
to steer a packet along a particular path. This path is encoded as an ordere rivedContent="RFC8402"/> is based on
d list of segments. source-routed tunnels to steer a packet along a particular path. This path
When applied to the MPLS dataplane <xref target="I-D.ietf-spring-segment- is encoded as an ordered list of segments. When applied to the MPLS data
routing-mpls"/>, each segment is an LSP (Label Switched Path) with an associated plane <xref target="RFC8660" format="default" sectionFormat="of" derivedConte
MPLS label value. nt="RFC8660"/>, each segment is an LSP
Hence, label stacking is used to represent the ordered list of segments a (Label Switched Path) with an associated MPLS label value. Hence, label
nd the label stack associated with an SR tunnel can be seen as nested LSPs (LSP stacking is used to represent the ordered list of segments, and the label
hierarchy) in the MPLS architecture. stack associated with an SR tunnel can be seen as nested LSPs (LSP
</t> hierarchy) in the MPLS architecture.
<t> </t>
<t pn="section-1-2">
Using label stacking to encode the list of segments has implications on t he label stack depth. Using label stacking to encode the list of segments has implications on t he label stack depth.
</t> </t>
<t pn="section-1-3">
<t> Traffic load-balancing over ECMP (Equal-Cost Multipath) or LAGs (Link
Traffic load-balancing over ECMP (Equal Cost Multi Path) or LAGs (Link Aggreg Aggregation Groups) is usually based on a hashing function. The local node
ation Groups) is usually based on a that performs the load-balancing is required to read some header fields in
hashing function. The local node which performs the load-balancing is require the incoming packets and then compute a hash based on those fields. The
d to read some header fields in the incoming packets result of the hash is finally mapped to a list of outgoing next hops. The
and then computes a hash based on those fields. The result of the hash is fin hashing technique is required to perform a per-flow load-balancing and
ally mapped to a list of outgoing nexthops. thus, prevents packet misordering. For IP traffic, the usual fields that
The hashing technique is required to perform a per-flow load-balancing and th are hashed are the source address, the destination address, the protocol
us prevents packet misordering. For IP traffic, the usual fields that are hashed type, and, if provided by the upper layer, the source port and destination
are port.
the source address, the destination address, the protocol type, and, if provi
ded by the upper layer, the source port and destination port.
</t> </t>
<t> <t pn="section-1-4">
The MPLS architecture brings some challenges when an LSR tries to look up at The MPLS architecture brings some challenges when an LSR (Label Switching
header fields. An LSR (Label Switching Router) needs be able to look up at heade Router) tries to look up at header fields. An LSR needs be able to look up
r fields that are beyond the MPLS label stack while the MPLS header does not pro at header fields that are beyond the MPLS label stack while the MPLS header
vide any information about the upper layer protocol. does not provide any information about the upper-layer protocol. An LSR
An LSR must perform a deeper inspection compared to an ingress router which c must perform a deeper inspection compared to an ingress router, which could
ould be challenging for some hardware. be challenging for some hardware. Entropy labels (ELs) <xref target="RFC6790
Entropy label (EL) <xref target="RFC6790"/> is a technique used in the MPLS d " format="default" sectionFormat="of" derivedContent="RFC6790"/> are used in the
ata MPLS data
plane to provide entropy for load-balancing. plane to provide entropy for load-balancing. The idea behind the entropy
The idea behind the entropy label is that the ingress router computes a hash label is that the ingress router computes a hash based on several fields
based on several fields from a given packet and places the result in an addition from a given packet and places the result in an additional label named
al label, named "entropy label". "entropy label". Then, this entropy label can be used as part of the hash
Then, this entropy label can be used as part of the hash keys used by an LSR. keys used by an LSR. Using the entropy label as part of the hash keys
Using the entropy label as part of the hash keys reduces the need for deep pack reduces the need for deep packet inspection in the LSR while keeping a good
et inspection in the LSR while keeping a good level of entropy in the load-balan level of entropy in the load-balancing. When the entropy label is used,
cing. the keys used in the hashing functions are still a local configuration
When the entropy label is used, the keys used in the hashing functions are st matter, and an LSR may use solely the entropy label or a combination of
ill a local configuration matter and an LSR may use solely the entropy label or multiple fields from the incoming packet.
a combination of multiple fields from the incoming packet. </t>
</t> <t pn="section-1-5">
<t>
When using LSP When using LSP
hierarchies, there are implications on how <xref target="RFC6790"/> should be hierarchies, there are implications on how <xref target="RFC6790" format="def ault" sectionFormat="of" derivedContent="RFC6790"/> should be
applied. The current document addresses the case where a hierarchy applied. The current document addresses the case where a hierarchy
is created at a single LSR as required by Segment Routing. is created at a single LSR as required by Segment Routing.
</t> </t>
<t> <t pn="section-1-6">
A use-case requiring load-balancing with SR is given in <xref target="usecase A use case requiring load-balancing with SR is given in <xref target="usecase
"/>. A recommended solution is " format="default" sectionFormat="of" derivedContent="Section 3"/>. A recommend
described in <xref target="solution"/> keeping in consideration the limitatio ed solution is
ns of described in <xref target="solution" format="default" sectionFormat="of" deri
implementations when applying <xref target="RFC6790"/> to deeper label stacks vedContent="Section 7"/> keeping in consideration the limitations of
. implementations when applying <xref target="RFC6790" format="default" section
Format="of" derivedContent="RFC6790"/> to deeper label stacks.
Options that were considered to arrive at the recommended solution Options that were considered to arrive at the recommended solution
are documented for historical purposes in <xref target="other-options"/>. are documented for historical purposes in <xref target="other-options" format
="default" sectionFormat="of" derivedContent="Section 10"/>.
</t>
<section title="Requirements Language" toc="default">
<t> The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in
BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/> when, and only when,
they appear in all
capitals, as shown here.
</t> </t>
<section toc="include" numbered="true" removeInRFC="false" pn="section-1.1
">
<name slugifiedName="name-requirements-language">Requirements Language</
name>
<t pn="section-1.1-1">
The key words "<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 "<bcp14>OPTIONAL</bcp14>" in this document are to
be
interpreted as described in BCP 14 <xref target="RFC2119" format="default" s
ectionFormat="of" derivedContent="RFC2119"/> <xref target="RFC8174" format="defa
ult" sectionFormat="of" derivedContent="RFC8174"/> when, and only when, they app
ear in all capitals, as
shown here.
</t>
</section> </section>
</section> </section>
<section toc="include" numbered="true" removeInRFC="false" pn="section-2">
<section title="Abbreviations and Terminology" toc="default"> <name slugifiedName="name-abbreviations-and-terminolo">Abbreviations and T
<t> erminology</name>
<list style="hanging"> <dl newline="false" spacing="normal" indent="10" pn="section-2-1">
<t>Adj-SID - Adjacency Segment Identifier</t> <dt pn="section-2-1.1">Adj-SID</dt>
<t>ECMP - Equal Cost Multi Path</t> <dd pn="section-2-1.2">Adjacency Segment Identifier</dd>
<t>EL - Entropy Label</t> <dt pn="section-2-1.3">ECMP</dt>
<t>ELI - Entropy Label Indicator</t> <dd pn="section-2-1.4">Equal-Cost Multipath</dd>
<t>ELC - Entropy Label Capability</t> <dt pn="section-2-1.5">EL</dt>
<t>ERLD - Entropy Readable Label Depth</t> <dd pn="section-2-1.6">Entropy Label</dd>
<t>FEC - Forwarding Equivalent Class</t> <dt pn="section-2-1.7">ELI</dt>
<t>LAG - Link Aggregation Group</t> <dd pn="section-2-1.8">Entropy Label Indicator</dd>
<t>LSP - Label Switched Path</t> <dt pn="section-2-1.9">ELC</dt>
<t>LSR - Label Switching Router</t> <dd pn="section-2-1.10">Entropy Label Capability</dd>
<t>MPLS - Multiprotocol Label Switching</t> <dt pn="section-2-1.11">ERLD</dt>
<t>MSD - Maximum SID Depth</t> <dd pn="section-2-1.12">Entropy Readable Label Depth</dd>
<t>Node-SID - Node Segment Identifier</t> <dt pn="section-2-1.13">FEC</dt>
<t>OAM - Operation, Administration and Maintenance</t> <dd pn="section-2-1.14">Forwarding Equivalence Class</dd>
<t>RLD - Readable Label Depth</t> <dt pn="section-2-1.15">LAG</dt>
<t>SID - Segment Identifier</t> <dd pn="section-2-1.16">Link Aggregation Group</dd>
<t>SPT - Shortest Path Tree</t> <dt pn="section-2-1.17">LSP</dt>
<t>SR - Segment Routing</t> <dd pn="section-2-1.18">Label Switched Path</dd>
<t>SRGB - Segment Routing Global Block</t> <dt pn="section-2-1.19">LSR</dt>
<t>VPN - Virtual Private Network</t> <dd pn="section-2-1.20">Label Switching Router</dd>
</list> <dt pn="section-2-1.21">MPLS</dt>
</t> <dd pn="section-2-1.22">Multiprotocol Label Switching</dd>
<dt pn="section-2-1.23">MSD</dt>
<dd pn="section-2-1.24">Maximum SID Depth</dd>
<dt pn="section-2-1.25">Node SID</dt>
<dd pn="section-2-1.26">Node Segment Identifier</dd>
<dt pn="section-2-1.27">OAM</dt>
<dd pn="section-2-1.28">Operations, Administration, and Maintenance</dd>
<dt pn="section-2-1.29">RLD</dt>
<dd pn="section-2-1.30">Readable Label Depth</dd>
<dt pn="section-2-1.31">SID</dt>
<dd pn="section-2-1.32">Segment Identifier</dd>
<dt pn="section-2-1.33">SPT</dt>
<dd pn="section-2-1.34">Shortest Path Tree</dd>
<dt pn="section-2-1.35">SR</dt>
<dd pn="section-2-1.36">Segment Routing</dd>
<dt pn="section-2-1.37">SRGB</dt>
<dd pn="section-2-1.38">Segment Routing Global Block</dd>
<dt pn="section-2-1.39">VPN</dt>
<dd pn="section-2-1.40">Virtual Private Network</dd>
</dl>
</section> </section>
<section anchor="usecase" title="Use-case requiring multipath load-balancing" <section anchor="usecase" toc="include" numbered="true" removeInRFC="false"
toc="default"> pn="section-3">
<figure title="Figure 1: Traffic engineering use-case"> <name slugifiedName="name-use-case-requiring-multipat">Use Case Requiring
<artwork> Multipath Load-Balancing</name>
+------+ <t pn="section-3-1">
| | Traffic engineering is one of the applications of MPLS and is also a
+-------| P3 |-----+ requirement for Segment Routing <xref target="RFC7855" format="default" sectionF
| +-----| |---+ | ormat="of" derivedContent="RFC7855"/>. Consider the
L3| |L4 +------+ L1| |L2 +----+ topology shown in <xref target="fig_TE_use_case" format="default" sectionFormat=
| | | | +--| P4 |--+ "of" derivedContent="Figure 1"/>. The LSR S requires data to be sent to LSR D a
+-----+ +-----+ +-----+ | +----+ | +-----+ long
| S |-----| P1 |------------| P2 |--+ +--| D | a traffic-engineered path that goes over the link L1. Good load-balancing is
| | | | | |--+ +--| | also required across equal-cost paths (including parallel links). To steer
+-----+ +-----+ +-----+ | +----+ | +-----+ traffic along a path that crosses link L1, the label stack that LSR S creates
+--| P5 |--+ consists of a label to the Node SID of LSR P3 stacked over the label for the
+----+ Adj-SID (Adjacency Segment Identifier) of link L1 and that in turn is stacked
S=Source LSR, D=Destination LSR, P1,P2,P3,P4,P5=Transit LSRs, over the label to the Node SID of LSR D. For simplicity, lets assume that all
L1,L2,L3,L4=Links LSRs use the same label space for Segment Routing (as a reminder, it is called
the SRGB, Segment Routing Global Block). Let L_N-Px denote the label to be
</artwork> used to reach the Node SID of LSR Px. Let L_A-Ln denote the label used for
</figure> the Adj-SID for link Ln. In our example, the LSR S must use the label stack
<t> &lt;L_N-P3, L_A-L1, L_N-D&gt;. However, to achieve good load-balancing over
Traffic-engineering is one of the applications of MPLS and is the equal-cost paths P2-P4-D, P2-P5-D, and the parallel links L3 and L4, a
also a requirement for Segment Routing <xref target="RFC7855"/>. mechanism such as entropy labels <xref target="RFC6790" format="default" section
Consider the topology shown in Figure 1. Format="of" derivedContent="RFC6790"/> should be adapted
The LSR S requires data to be sent to LSR D along a traffic-engineered path t for Segment Routing. Indeed, the Source Packet Routing in Networking (SPRING)
hat goes over the link L1. architecture with the MPLS data plane <xref target="RFC8660" format="default" se
Good load-balancing is ctionFormat="of" derivedContent="RFC8660"/> uses nested
also required across equal cost paths (including parallel links). To MPLS LSPs composing the source-routed label stack.
steer traffic along a path that crosses link L1, the label stack </t>
that LSR S creates consists of a label to the Node-SID of LSR P3, <figure anchor="fig_TE_use_case" align="left" suppress-title="false" pn="f
stacked over the label for the Adj-SID (Adjacency Segment Identifier) of link igure-1">
L1 and that in <name slugifiedName="name-traffic-engineering-use-cas">Traffic-Engineeri
turn is stacked over the label to the Node-SID of LSR D. For ng Use Case</name>
simplicity lets assume that all LSRs use the same label space for <artwork name="" type="" align="left" alt="" pn="section-3-2.1">
Segment Routing (as a reminder, it is called the SRGB, Segment Routing Global +------+
Block). Let L_N-Px denote the label to be used | |
to reach the Node-SID of LSR Px. Let L_A-Ln denote the label used +-------| P3 |-----+
for the Adj-SID for link Ln. In our example, the LSR S must use the label | +-----| |---+ |
stack &lt;L_N-P3, L_A-L1, L_N-D&gt;. However, to L3| |L4 +------+ L1| |L2 +----+
achieve a good load-balancing over the equal cost paths P2-P4-D, | | | | +--| P4 |--+
P2-P5-D and the parallel links L3 and L4, a mechanism such as entropy +-----+ +-----+ +-----+ | +----+ | +-----+
labels <xref target="RFC6790"/> should be adapted for Segment Routing. | S |-----| P1 |------------| P2 |--+ +--| D |
Indeed, the SPRING architecture with the MPLS dataplane (<xref target="I-D.ie | | | | | |--+ +--| |
tf-spring-segment-routing-mpls"/>) uses nested MPLS LSPs composing the source ro +-----+ +-----+ +-----+ | +----+ | +-----+
uted label stack. +--| P5 |--+
</t> +----+
<t> Key:
S = Source LSR
D = Destination LSR
P1, P2, P3, P4, P5 = Transit LSRs
L1, L2, L3, L4 = Links
</artwork>
</figure>
<t pn="section-3-3">
An MPLS node may have limitations in the number of labels it can push. It may also have a limitation in the number of labels it can inspect when looking for hash keys during load-balancing. An MPLS node may have limitations in the number of labels it can push. It may also have a limitation in the number of labels it can inspect when looking for hash keys during load-balancing.
While the entropy label is normally inserted at the bottom of the transport t unnel, this may prevent an LSR from taking into account the EL in its load-balan cing function if the EL is too deep in the stack. While the entropy label is normally inserted at the bottom of the transport t unnel, this may prevent an LSR from taking into account the EL in its load-balan cing function if the EL is too deep in the stack.
In a Segment Routing environment, it is important to define the consideration s that needs to be taken into account when inserting an EL. In a Segment Routing environment, it is important to define the consideration s that need to be taken into account when inserting an EL.
Multiple ways to apply entropy labels were considered and are Multiple ways to apply entropy labels were considered and are
documented in <xref target="other-options"/> along with their trade-offs. A documented in <xref target="other-options" format="default" sectionFormat="of
recommended " derivedContent="Section 10"/> along with their trade-offs. A recommended
solution is described in <xref target="solution"/>. solution is described in <xref target="solution" format="default" sectionForm
</t> at="of" derivedContent="Section 7"/>.
</t>
</section> </section>
<section anchor="erld_definition" title="Entropy Readable Label Depth"> <section anchor="erld_definition" numbered="true" toc="include" removeInRFC=
<t> "false" pn="section-4">
<name slugifiedName="name-entropy-readable-label-dept">Entropy Readable La
bel Depth</name>
<t pn="section-4-1">
The Entropy Readable Label Depth (ERLD) is defined as the number of labels a router can both: The Entropy Readable Label Depth (ERLD) is defined as the number of labels a router can both:
<list style="letters"> </t>
<t>Read in an MPLS packet received on its incoming interface(s) (starting fro <ol spacing="normal" type="a" start="1" pn="section-4-2">
m the top of the stack).</t> <li pn="section-4-2.1" derivedCounter="a.">Read in an MPLS packet receiv
<t>Use in its load-balancing function.</t> ed on its incoming interface(s) (starting from the top of the stack).</li>
</list> <li pn="section-4-2.2" derivedCounter="b.">Use in its load-balancing fun
</t> ction.</li>
<t>The ERLD means that the router will perform load-balancing using the EL la </ol>
bel if the EL is placed within the first ERLD labels.</t> <t pn="section-4-3">The ERLD means that the router will perform load-balan
<t>A router capable of reading N labels but not using an EL located within th cing using the EL if the EL is placed within the first ERLD labels.</t>
ose N labels MUST consider its ERLD to be 0.</t> <t pn="section-4-4">A router capable of reading N labels but not using an
<t> EL located within those N labels <bcp14>MUST</bcp14> consider its ERLD to be 0.<
In a distributed switching architecture, each linecard may have a different c /t>
apability in terms of ERLD. For simplicity, an implementation MAY use the minimu <t pn="section-4-5">
m ERLD of all linecards as the ERLD value for the system. In a distributed switching architecture, each line card may have a
</t> different capability in terms of ERLD. For simplicity, an implementation
<t>There may also be a case where a router has a fast switching path (handled <bcp14>MAY</bcp14> use the minimum ERLD of all line cards as the ERLD value f
by an ASIC or network processor) and a slow switching path (handled by a CPU) w or the system.
ith a different ERLD for each switching path. Again, for simplicity's sake, an i </t>
mplementation MAY use the minimum ERLD as the ERLD value for the system.</t> <t pn="section-4-6">There may also be a case where a router has a fast swi
<t>The drawback of using a single ERLD for a system lower than the capability tching path
of one or more specific component is that it may increase the number of ELI/ELs (handled by an Application-Specific Integrated Circuit, or ASIC, or network p
inserted. This leads to an increase of the label stack size and may have an imp rocessor) and a slow switching path (handled by a CPU) with a different ERLD for
act on the capability of the ingress node to push this label stack.</t> each switching path. Again, for simplicity's sake, an implementation <bcp14>MAY
<t>Examples:</t> </bcp14> use the minimum ERLD as the ERLD value for the system.</t>
<figure title="Figure 2: Label stacks with ELI/EL"> <t pn="section-4-7">The drawback of using a single ERLD for a system lower
<artwork> than the capability of one or more specific components is that it may increase
| Payload | the number of ELI/ELs inserted. This leads to an increase of the label stack siz
+----------+ e and may have an impact on the capability of the ingress node to push this labe
| Payload | | EL | P7 l stack.</t>
+----------+ +----------+ <t pn="section-4-8">Examples:</t>
| Payload | | EL | | ELI | <figure anchor="fig_label_stacks" align="left" suppress-title="false" pn="
+----------+ +----------+ +----------+ figure-2">
| Payload | | EL | | ELI | | Label 50 | <name slugifiedName="name-label-stacks-with-eli-el">Label Stacks with EL
+----------+ +----------+ +----------+ +----------+ I/EL</name>
| Payload | | EL | | ELI | | Label 40 | | Label 40 | <artwork name="" type="" align="left" alt="" pn="section-4-9.1">
+----------+ +----------+ +----------+ +----------+ +----------+ | Payload |
| EL | | ELI | | Label 30 | | Label 30 | | Label 30 | +----------+
+----------+ +----------+ +----------+ +----------+ +----------+ | Payload | | EL | P7
| ELI | | Label 20 | | Label 20 | | Label 20 | | Label 20 | +----------+ +----------+
+----------+ +----------+ +----------+ +----------+ +----------+ | Payload | | EL | | ELI |
| Label 16 | | Label 16 | | Label 16 | | Label 16 | | Label 16 | P1 +----------+ +----------+ +----------+
+----------+ +----------+ +----------+ +----------+ +----------+ | Payload | | EL | | ELI | | Label 50 |
Packet 1 Packet 2 Packet 3 Packet 4 Packet 5 +----------+ +----------+ +----------+ +----------+
</artwork> | Payload | | EL | | ELI | | Label 40 | | Label 40 |
</figure> +----------+ +----------+ +----------+ +----------+ +----------+
<t> | EL | | ELI | | Label 30 | | Label 30 | | Label 30 |
In Figure 2, we consider the displayed packets received on a router interface +----------+ +----------+ +----------+ +----------+ +----------+
. We consider also a single ERLD value for the router. | ELI | | Label 20 | | Label 20 | | Label 20 | | Label 20 |
<list style="symbols"> +----------+ +----------+ +----------+ +----------+ +----------+
<t>If the router has an ERLD of 3, it will be able to load-balance Packet 1 d | Label 16 | | Label 16 | | Label 16 | | Label 16 | | Label 16 | P1
isplayed in Figure 2 using the EL as part of the load-balancing keys. The ERLD v +----------+ +----------+ +----------+ +----------+ +----------+
alue of 3 means that the router can read and take into account the entropy label Packet 1 Packet 2 Packet 3 Packet 4 Packet 5
for load-balancing if it is placed between position 1 (top of the MPLS label st </artwork>
ack) and position 3.</t> </figure>
<t>If the router has an ERLD of 5, it will be able to load-balance Packets 1 <t pn="section-4-10">
to 3 in Figure 2 using the EL as part of the load-balancing keys. Packets 4 and In <xref target="fig_label_stacks" format="default" sectionFormat="of" derive
5 have the EL placed at a position greater than 5, so the router is not able to dContent="Figure 2"/>, we consider the displayed packets received on a router in
read it and use as part of the load-balancing keys.</t> terface. We consider also a single ERLD value for the router.
<t>If the router has an ERLD of 10, it will be able to load-balance all the p </t>
ackets displayed in Figure 2 using the EL as part of the load-balancing keys.</t <ul spacing="normal" bare="false" empty="false" pn="section-4-11">
> <li pn="section-4-11.1">If the router has an ERLD of 3, it will be able
</list> to load-balance Packet 1 displayed in <xref target="fig_label_stacks" format="de
</t> fault" sectionFormat="of" derivedContent="Figure 2"/> using the EL as part of th
e load-balancing keys. The ERLD value of 3 means that the router can read and ta
<t>To allow an efficient load-balancing based on entropy labels, a router run ke into account the entropy label for load-balancing if it is placed between pos
ning SPRING SHOULD advertise its ERLD (or ERLDs), so all the other SPRING router ition 1 (top of the MPLS label stack) and position 3.</li>
s in the network are aware of its capability. How this advertisement is done is <li pn="section-4-11.2">If the router has an ERLD of 5, it will be able
outside the scope of this document (see <xref target="erld"/> for potential appr to load-balance Packets
oaches). 1 to 3 in <xref target="fig_label_stacks" format="default" sectionFormat="of"
</t> derivedContent="Figure 2"/> using the EL as part of the load-balancing keys. Pa
<t> ckets
4 and 5 have the EL placed at a position greater than 5, so the router is
not able to read it and use it as part of the load-balancing keys.</li>
<li pn="section-4-11.3">If the router has an ERLD of 10, it will be able
to load-balance all the packets displayed in <xref target="fig_label_stacks" fo
rmat="default" sectionFormat="of" derivedContent="Figure 2"/> using the EL as pa
rt of the load-balancing keys.</li>
</ul>
<t pn="section-4-12">To allow an efficient load-balancing based on entropy
labels, a router running SPRING <bcp14>SHOULD</bcp14> advertise its ERLD (or ER
LDs), so all the other SPRING routers in the network are aware of its capability
. How this advertisement is done is outside the scope of this document (see <xre
f target="erld" format="default" sectionFormat="of" derivedContent="Section 7.2.
1"/> for potential approaches).
</t>
<t pn="section-4-13">
To advertise an ERLD value, a SPRING router: To advertise an ERLD value, a SPRING router:
<list style="symbols"> </t>
<t>MUST be entropy label capable and, as a consequence, MUST apply the datapl <ul spacing="normal" bare="false" empty="false" pn="section-4-14">
ane procedures defined in <xref target="RFC6790"/>.</t> <li pn="section-4-14.1">
<t>MUST be able to read an ELI/EL which is located within its ERLD value.</t> <bcp14>MUST</bcp14> be entropy label capable and, as a consequence, <b
<t>MUST take into account an EL within the first ERLD labels in its load-bala cp14>MUST</bcp14> apply the data-plane procedures defined in <xref target="RFC67
ncing function.</t> 90" format="default" sectionFormat="of" derivedContent="RFC6790"/>.</li>
</list> <li pn="section-4-14.2">
</t> <bcp14>MUST</bcp14> be able to read an ELI/EL, which is located within
</section> its ERLD value.</li>
<section anchor="msd" title="Maximum SID Depth"> <li pn="section-4-14.3">
<t> <bcp14>MUST</bcp14> take into account an EL within the first ERLD labe
The Maximum SID Depth defines the maximum number of labels that a particular ls in its load-balancing function.</li>
node can impose on a packet. This can include any kind of labels (service, entro </ul>
py, transport...). </section>
In an MPLS network, the MSD is a limit of the head-end of an SR tunnel or a B <section anchor="msd" numbered="true" toc="include" removeInRFC="false" pn="
inding-SID section-5">
anchor node that performs imposition of additional labels on an existing labe <name slugifiedName="name-maximum-sid-depth">Maximum SID Depth</name>
l stack. <t pn="section-5-1">
</t> The Maximum SID Depth defines the maximum number of labels that a
<t> particular node can impose on a packet. This can include any kind of labels
Depending on the number of MPLS operations (POP, SWAP...) to be performed bef (service, entropy, transport, etc.). In an MPLS network, the MSD is a
ore the PUSH, the MSD can vary due to hardware or software limitations. limit of the head-end of an SR tunnel or a Binding SID anchor node that
As for the ERLD, different MSD limits can exist within a single node based on performs imposition of additional labels on an existing label stack.
the linecard types used in a distributed switching system. Thus, the MSD is a p </t>
er link and/or per node property. <t pn="section-5-2">
</t> Depending on the number of MPLS operations (POP, SWAP, etc.) to be performed
<t> before the PUSH, the MSD can vary due to hardware or software limitations.
An external controller can be used to program a label stack on a As for the ERLD, different MSD limits can exist within a single node based
particular node. This node SHOULD advertise its MSD to the controller in orde on the line-card types used in a distributed switching system. Thus, the MSD
r to let the controller know the maximum label stack depth of the path computed is a per link and/or per-node property.
that is supported on the head-end. </t>
How this advertisement is done is <t pn="section-5-3">
outside the scope of this document (<xref target="I-D.ietf-isis-segment-routi An external controller can be used to program a label stack on a particular
ng-msd"/>, <xref target="I-D.ietf-isis-segment-routing-msd"/> and <xref target=" node. This node <bcp14>SHOULD</bcp14> advertise its MSD to the controller
I-D.ietf-idr-bgp-ls-segment-routing-msd"/> provide examples of advertisement of in order to let the controller know the maximum label stack depth of the
MSD). path computed that is supported on the head-end.
As the controller does not have
the knowledge of the entire label stack to be pushed by the node, in addition How this advertisement is done is outside the scope of this
to the MSD value, the document. (<xref target="RFC8476" format="default" sectionFormat="of" derived
node SHOULD advertise the type of the MSD. Content="RFC8476"/>, <xref target="RFC8491" format="default" sectionFormat="of"
For instance, the MSD value can represent the limit for pushing transport lab derivedContent="RFC8491"/>, and <xref target="I-D.ietf-idr-bgp-ls-segment-routin
els only while in reality the node can push an additional service label. As anot g-msd" format="default" sectionFormat="of" derivedContent="MSD-BGP"/> provide
her example, the MSD value can represent the full limit of the node including al examples of advertisement of the MSD.) As the controller does not have the
l label types (transport, service, entropy...). knowledge of the entire label stack to be pushed by the node, in addition
This gives the ability for the controller to program a label stack while leav to the MSD value, the node <bcp14>SHOULD</bcp14> advertise the type of the
ing room for the local node to add more labels (e.g., service, entropy,...) with MSD. For instance, the MSD value can represent the limit for pushing
out reaching the hardware/software limit. transport labels only while in reality the node can push an additional
If the node does not provide the meaning of the MSD value, the controller cou service label. As another example, the MSD value can represent the full
ld program an LSP using a number of labels equal to the full limit of the node. limit of the node including all label types (transport, service, entropy,
When receiving this label stack from the controller, the ingress node may not be etc.). This gives the ability for the controller to program a label stack
able to add any service (L2VPN, L3VPN, EVPN...) label on top of this label stac while leaving room for the local node to add more labels (e.g., service,
k. entropy, etc.) without reaching the hardware/software limit. If the node
The consequence could be for the ingress node to drop service packets that sh does not provide the meaning of the MSD value, the controller could program
ould have been forwarded over the LSP. an LSP using a number of labels equal to the full limit of the node. When
</t> receiving this label stack from the controller, the ingress node may not be
<figure title="Figure 3"> able to add any service (L2VPN, L3VPN, EVPN, etc.) label on top of this
<artwork> label stack. The consequence could be for the ingress node to drop service
packets that should have been forwarded over the LSP.
</t>
<figure anchor="fig_packet" align="left" suppress-title="false" pn="figure
-3">
<name slugifiedName="name-topology-illustrating-label">Topology Illustra
ting Label Stack Reduction</name>
<artwork name="" type="" align="left" alt="" pn="section-5-4.1">
P7 ---- P8 ---- P9 P7 ---- P8 ---- P9
/ \ / \
PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- PE2 PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- PE2
| \ | | \ |
----&gt; P10 \ |
IP Pkt | \ | IP Pkt | \ |
P11 --- P12 --- P13 P11 --- P12 --- P13
100 10000 100 10000
</artwork> </artwork>
</figure> </figure>
<t> <t pn="section-5-5">
In Figure 3, an IP packet comes into the MPLS network at PE1. All metrics are In <xref target="fig_packet" format="default" sectionFormat="of" derivedConte
considered equal to 1 except P12-P13 which is 10000 and P11-P12 which is 100. nt="Figure 3"/>, an IP packet comes into the MPLS network at PE1. All metrics
PE1 wants to steer the traffic using a SPRING path to PE2 along PE1->P1->P7-> are considered equal to 1 except P12-P13, which is 10000, and P11-P12,
P8->P9->P4->P5->P10->P11->P12->P13->PE2. which is 100. PE1 wants to steer the traffic using a SPRING path to PE2
By using Adj-SIDs only, PE1 (acting as an I-LSR) will be required to push 10 along PE1 -&gt; P1 -&gt; P7 -&gt; P8 -&gt; P9 -&gt; P4 -&gt; P5 -&gt; P10 -&g
labels on the IP packet received and thus requires an MSD of 10. t; P11 -&gt; P12 -&gt; P13
If the IP packet should be carried over an MPLS service like a regular layer -&gt; PE2. By using Adj-SIDs only, PE1 (acting as an ingress LSR, also known
3 VPN, an additional service label should be imposed, requiring an MSD of 11 for as an I-LSR) will be required to push 10 labels on the IP packet received
PE1. and thus, requires an MSD of 10. If the IP packet should be carried over
In addition, if PE1 wants to insert an ELI/EL for load-balancing purpose, PE1 an MPLS service like a regular layer 3 VPN, an additional service label
will need to push 13 labels on the IP packet requiring an MSD of 13. should be imposed requiring an MSD of 11 for PE1. In addition, if PE1
</t> wants to insert an ELI/EL for load-balancing purposes, PE1 will need to
<t> push 13 labels on the IP packet requiring an MSD of 13.
In the SPRING architecture, Node-SIDs or Binding-SIDs can be used to reduce t </t>
he label stack size. As an example, to steer the traffic on the same path as bef <t pn="section-5-6">
ore, PE1 could use the following label stack: &lt;Node_P9, Node_P5, Binding_P5, In the SPRING architecture, Node SIDs or Binding SIDs can be used to reduce t
Node_PE2&gt;. he label stack size. As an example, to steer the traffic on the same path as bef
In this example we consider a combination of Node-SIDs and a Binding-SID adve ore, PE1 could use the following label stack: &lt;Node_P9, Node_P5, Binding_P5,
rtised by P5 that will stitch the traffic along the path P10->P11->P12->P13. The Node_PE2&gt;.
instruction associated with the Binding-SID at P5 is thus to swap Binding_P5 to In this example, we consider a combination of Node SIDs and a Binding SID
Adj_P12-P13 and then push &lt;Adj_P11-P12, Node_P11&gt;. advertised by P5 that will stitch the traffic along the path P10 -&gt; P11
P5 acts as a stitching node that pushes additional labels on an existing labe -&gt; P12 -&gt; P13. The instruction associated with the Binding SID at P5 is
l stack, P5's MSD needs also to be taken into account and may limit the number o thus to swap Binding_P5 to Adj_P12-P13 and then push &lt;Adj_P11-P12, Node_P11&
f labels that can be imposed. gt;.
</t> P5 acts as a stitching node that pushes additional labels on an existing labe
</section> l stack; P5's MSD needs also to be taken into account and may limit the number o
<section anchor="stitching" title="LSP stitching using the Binding-SID"> f labels that can be imposed.
<t> </t>
The Binding-SID allows binding a segment identifier to an existing LSP. As ex </section>
amples, the Binding-SID can represent an RSVP-TE tunnel, an LDP path (through th <section anchor="stitching" numbered="true" toc="include" removeInRFC="false
e mapping server advertisement), or a SPRING path. " pn="section-6">
Each tail-end router of an MPLS LSP associated with a Binding-SID has its own <name slugifiedName="name-lsp-stitching-using-the-bin">LSP Stitching Using
entropy label capability. The entropy label capability of the associated LSP is the Binding SID</name>
advertised in the control plane protocol used to signal the LSP. <t pn="section-6-1">
</t> The Binding SID allows binding a segment identifier to an existing LSP. As
<t> examples, the Binding SID can represent an RSVP-TE tunnel, an LDP path
In Figure 4, we consider that: (through the Mapping Server Advertisement), or a SPRING path. Each
<list style="symbols"> tail-end router of an MPLS LSP associated with a Binding SID has its own
<t>P6, PE2, P10, P11, P12, P13 are pure LDP routers.</t> entropy label capability. The entropy label capability of the associated
<t>PE1, P1, P2, P3, P4, P7, P8, P9 are pure SPRING routers.</t> LSP is advertised in the control-plane protocol used to signal the LSP.
<t>P5 is running SPRING and LDP.</t> </t>
<t>P5 acts as a mapping server and advertises Prefix SIDs for the LDP FEC <t pn="section-6-2">
s: an index value of 20 is used for PE2.</t> In <xref target="fig_stitching_example" format="default" sectionFormat="of" deri
<t>All SPRING routers use an SRGB of [1000, 1999].</t> vedContent="Figure 4"/>, we consider that:
<t>P6 advertises label 20 for the PE2 FEC.</t> </t>
<t>Traffic from PE1 to PE2 uses the shortest path.</t> <ul spacing="normal" bare="false" empty="false" pn="section-6-3">
</list> <li pn="section-6-3.1">P6, PE2, P10, P11, P12, and P13 are pure LDP rout
</t> ers.</li>
<figure> <li pn="section-6-3.2">PE1, P1, P2, P3, P4, P7, P8, and P9 are pure SPRI
<artwork> NG routers.</li>
<li pn="section-6-3.3">P5 is running SPRING and LDP.</li>
<li pn="section-6-3.4">P5 acts as a Mapping Server and advertises Prefix
-SIDs for the LDP FECs: an index value of 20 is used for PE2.</li>
<li pn="section-6-3.5">All SPRING routers use an SRGB of [1000, 1999].</
li>
<li pn="section-6-3.6">P6 advertises label 20 for the PE2 FEC.</li>
<li pn="section-6-3.7">Traffic from PE1 to PE2 uses the shortest path.</
li>
</ul>
<figure anchor="fig_stitching_example" align="left" suppress-title="false"
pn="figure-4">
<name slugifiedName="name-example-illustrating-need-f">Example Illustrat
ing Need for ELC Propagation</name>
<artwork name="" type="" align="left" alt="" pn="section-6-4.1">
PE1 ----- P1 -- P2 -- P3 -- P4 ---- P5 --- P6 --- PE2 PE1 ----- P1 -- P2 -- P3 -- P4 ---- P5 --- P6 --- PE2
--&gt; +----+ +----+ +----+ +----+
--> +----+ +----+ +----+ +----+
IP Pkt | IP | | IP | | IP | | IP | IP Pkt | IP | | IP | | IP | | IP |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
|1020| |1020| | 20 | |1020| |1020| | 20 |
+----+ +----+ +----+ +----+ +----+ +----+
SPRING LDP SPRING LDP
</artwork> </artwork>
</figure> </figure>
<t>In terms of packet forwarding, by learning the mapping-server advertis <t pn="section-6-5">In terms of packet forwarding, by learning the Mapping
ement from P5, PE1 imposes a label 1020 to an IP packet destined to PE2. Server Advertisement from P5, PE1 imposes a label 1020 to an IP packet destined
SPRING routers along the shortest path to PE2 will switch the traffic unt to PE2.
il it reaches P5. P5 will perform the LSP stitching by swapping the SPRING label SPRING routers along the shortest path to PE2 will switch the traffic
1020 to the LDP label 20 advertised by the nexthop P6. until it reaches P5. P5 will perform the LSP stitching by swapping the
SPRING label 1020 to the LDP label 20 advertised by the next hop P6.
P6 will finally forward the packet using the LDP label towards PE2.</t> P6 will finally forward the packet using the LDP label towards PE2.</t>
<t> <t pn="section-6-6">
PE1 cannot push an ELI/EL for the Binding-SID without knowing that the ta PE1 cannot push an ELI/EL for the Binding SID without knowing that the
il-end of the LSP associated with the binding (PE2) is entropy label capable. tail end of the LSP associated with the binding (PE2) is entropy label ca
</t> pable.
<t> </t>
To accommodate the mix of signaling protocols involved during the stitchi <t pn="section-6-7">
ng, the entropy label capability SHOULD be propagated between the signaling doma To accommodate the mix of signaling protocols involved during the stitchi
ins. ng, the entropy label capability <bcp14>SHOULD</bcp14> be propagated between the
Each Binding-SID SHOULD have its own entropy label capability that MUST b signaling domains.
e inherited from the entropy label capability of the associated LSP. Each Binding SID <bcp14>SHOULD</bcp14> have its own entropy label capabil
If the router advertising the Binding-SID does not know the ELC state of ity that <bcp14>MUST</bcp14> be inherited from the entropy label capability of t
the target FEC, it MUST NOT set the ELC for the Binding-SID. he associated LSP.
An ingress node MUST NOT push an ELI/EL associated with a Binding-SID unl If the router advertising the Binding SID does not know the ELC state
ess this Binding-SID has the entropy label capability. of the target FEC, it <bcp14>MUST NOT</bcp14> set the ELC for the
How the entropy label capability is advertised for a Binding-SID is outsi Binding SID.
de the scope of this document (see <xref target="erld"/> for potential approache An ingress node <bcp14>MUST NOT</bcp14> push an ELI/EL associated with
s). a Binding SID unless this Binding SID has the entropy label capability.
</t> How the entropy label capability is advertised for a Binding SID is outsi
<t> de the scope of this document (see <xref target="erld" format="default" sectionF
In our example, if PE2 is LDP entropy label capable, it will add the entr ormat="of" derivedContent="Section 7.2.1"/> for potential approaches).
opy label capability in its LDP advertisement. When P5 receives the FEC/label bi </t>
nding for PE2, it learns about the ELC and can set the ELC in the mapping server <t pn="section-6-8">
advertisement. Thus PE1 learns about the ELC of PE2 and may push an ELI/EL asso In our example, if PE2 is LDP entropy label capable, it will add the
ciated with the Binding-SID. entropy label capability in its LDP advertisement. When P5 receives
</t> the FEC/label binding for PE2, it learns about the ELC and can set the
<t> ELC in the Mapping Server Advertisement. Thus, PE1 learns about the
The proposed solution only works if the SPRING router advertising the Bin ELC of PE2 and may push an ELI/EL associated with the Binding SID.
ding-SID is also performing the dataplane LSP stitching. </t>
In our example, if the mapping server function is hosted on P8 instead of <t pn="section-6-9">
P5, P8 does not know about the ELC state of PE2's LDP FEC. As a consequence, it The proposed solution only works if the SPRING router advertising the
does not set the ELC for the associated Binding-SID. Binding SID is also performing the data-plane LSP stitching.
</t> In our example, if the Mapping Server function is hosted on P8 instead
</section> of P5, P8 does not know about the ELC state of PE2's LDP FEC. As a
consequence, it does not set the ELC for the associated Binding SID.
<section anchor="solution" title="Insertion of entropy labels for SPRING path" </t>
toc="default"> </section>
<section anchor="overview" title="Overview"> <section anchor="solution" toc="include" numbered="true" removeInRFC="false"
<t> pn="section-7">
The solution described in this section follows the dataplane processing d <name slugifiedName="name-insertion-of-entropy-labels">Insertion of Entrop
efined in <xref target="RFC6790"/>. Within a SPRING path, a node may be ingress, y Labels for SPRING Path</name>
egress, transit (regarding the entropy label processing described in <xref targ <section anchor="overview" numbered="true" toc="include" removeInRFC="fals
et="RFC6790"/>), or it can be any combination of those. e" pn="section-7.1">
<name slugifiedName="name-overview">Overview</name>
<t pn="section-7.1-1">
The solution described in this section follows the data-plane processing
defined in <xref target="RFC6790" format="default" sectionFormat="of" derivedCon
tent="RFC6790"/>. Within a SPRING path, a node may be ingress, egress, transit (
regarding the entropy label processing described in <xref target="RFC6790" forma
t="default" sectionFormat="of" derivedContent="RFC6790"/>), or it can be any com
bination of those.
For example: For example:
<list style="symbols"> </t>
<t>The ingress node of a SPRING domain can be an ingress node fro <ul spacing="normal" bare="false" empty="false" pn="section-7.1-2">
m an entropy label perspective.</t> <li pn="section-7.1-2.1">The ingress node of a SPRING domain can be an
<t>Any LSR terminating a segment of the SPRING path is an egress ingress node from an entropy label perspective.</li>
node (because it terminates the segment) but can also be a transit node if the S <li pn="section-7.1-2.2">Any LSR terminating a segment of the SPRING p
PRING path is not terminated because there is a subsequent SPRING MPLS label in ath is an egress node (because it terminates the segment) but can also be a tran
the stack.</t> sit node if the SPRING path is not terminated because there is a subsequent SPRI
<t>Any LSR processing a Binding-SID may be a transit node and an NG MPLS label in the stack.</li>
ingress node (because it may push additional labels when processing the Binding- <li pn="section-7.1-2.3">Any LSR processing a Binding SID may be a tra
SID).</t> nsit node and an
</list> ingress node (because it may push additional labels when processing
</t> the Binding SID).</li>
<t> </ul>
<t pn="section-7.1-3">
As described earlier, an LSR may have a limitation (the ERLD) on the dept h of the label stack that it As described earlier, an LSR may have a limitation (the ERLD) on the dept h of the label stack that it
can read and process in order to do multipath load-balancing based on entropy labels.</t> can read and process in order to do multipath load-balancing based on entropy labels.</t>
<t>If an EL does not occur within the ERLD of an <t pn="section-7.1-4">If an EL does not occur within the ERLD of an
LSR in the label stack of an MPLS packet that it receives, then it LSR in the label stack of an MPLS packet that it receives, then it
would lead to poor load-balancing at that LSR. Hence an ELI/EL pair would lead to poor load-balancing at that LSR. Hence, an ELI/EL pair
must be within the ERLD of the LSR in order for the LSR to use the EL must be within the ERLD of the LSR in order for the LSR to use the EL
during load-balancing. during load-balancing.
</t> </t>
<t> <t pn="section-7.1-5">
Adding a single ELI/EL pair for the entire SPRING path can also lead Adding a single ELI/EL pair for the entire SPRING path can also lead
to poor load-balancing as well because the ELI/EL may not occur within to poor load-balancing as well because the ELI/EL may not occur within
the ERLD of some LSR on the path (if too deep) or may not be present the ERLD of some LSR on the path (if too deep) or may not be present
in the stack when it reaches some LSRs (if it is too shallow). in the stack when it reaches some LSRs (if it is too shallow).
</t> </t>
<t> <t pn="section-7.1-6">
In order for the EL to occur within the ERLD of LSRs along the path In order for the EL to occur within the ERLD of LSRs along the path
corresponding to a SPRING label stack, multiple &lt;ELI, EL&gt; pairs MAY be corresponding to a SPRING label stack, multiple &lt;ELI, EL&gt; pairs <bcp14> MAY</bcp14> be
inserted in this label stack. inserted in this label stack.
</t> </t>
<t> <t pn="section-7.1-7">
The insertion of an ELI/EL MUST occur only with a SPRING label advertised by The insertion of an ELI/EL <bcp14>MUST</bcp14> occur only with a SPRING
an LSR that advertised an ERLD (the LSR is entropy label capable) or with a SPRI label advertised by an LSR that advertised an ERLD (the LSR is entropy
NG label associated with a Binding-SID that has the ELC set. label capable) or with a SPRING label associated with a Binding SID that has
</t> the ELC set.
<t> </t>
<t pn="section-7.1-8">
The ELs among multiple &lt;ELI, EL&gt; pairs inserted in the The ELs among multiple &lt;ELI, EL&gt; pairs inserted in the
stack MAY be the same or different. The LSR that inserts &lt;ELI, EL&gt; pair s stack <bcp14>MAY</bcp14> be the same or different. The LSR that inserts &lt;E LI, EL&gt; pairs
can have limitations on the number of such pairs that it can insert can have limitations on the number of such pairs that it can insert
and also the depth at which it can insert them. If, due to and also the depth at which it can insert them. If, due to
limitations, the inserted ELs are at positions such that an LSR along limitations, the inserted ELs are at positions such that an LSR along
the path receives an MPLS packet without an EL in the label stack the path receives an MPLS packet without an EL in the label stack
within that LSR's ERLD, then the load-balancing performed by that LSR within that LSR's ERLD, then the load-balancing performed by that LSR
would be poor. An implementation MAY consider multiple criteria when insertin would be poor. An implementation <bcp14>MAY</bcp14> consider multiple criteri
g &lt;ELI, EL&gt; pairs. a when inserting &lt;ELI, EL&gt; pairs.
</t> </t>
<section anchor="ex1" title="Example 1 where the ingress node has a suffi <section anchor="ex1" numbered="true" toc="include" removeInRFC="false"
cient MSD"> pn="section-7.1.1">
<figure title="Figure 5"> <name slugifiedName="name-example-1-the-ingress-node-">Example 1: The
<artwork> Ingress Node Has a Sufficient MSD</name>
<figure anchor="fig_ex1" align="left" suppress-title="false" pn="figur
e-5">
<name slugifiedName="name-accommodating-msd-limitatio">Accommodating
MSD Limitations</name>
<artwork name="" type="" align="left" alt="" pn="section-7.1.1-1.1">
ECMP LAG LAG ECMP LAG LAG
PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- PE2 PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- PE2
</artwork>
</artwork> </figure>
</figure> <t pn="section-7.1.1-2">
<t> In <xref target="fig_ex1" format="default" sectionFormat="of" derivedCont
In Figure 5, PE1 wants to forward some MPLS VPN traffic over an explicit ent="Figure 5"/>, PE1 wants to forward some MPLS VPN traffic over an explicit pa
path to PE2 resulting in the following label stack to be pushed onto the receive th to PE2 resulting in the following label stack to be pushed onto the received
d IP header: &lt;Adj_P1P2, Adj_set_P2P3, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_P6PE2 IP header: &lt;Adj_P1P2, Adj_set_P2P3, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_P6PE2,
, VPN_label&gt;. VPN_label&gt;.
PE1 is limited to push a maximum of 11 labels (MSD=11). P2, P3 and P6 have an ER PE1 is limited to push a maximum of 11 labels (MSD=11). P2, P3, and P6 have an E
LD of 3 while others have an ERLD of 10. RLD of 3 while others have an ERLD of 10.
</t> </t>
<t> <t pn="section-7.1.1-3">
PE1 can only add two ELI/EL pairs in the label stack due to its MSD limit ation. It should insert them strategically to benefit load-balancing along the l ongest part of the path. PE1 can only add two ELI/EL pairs in the label stack due to its MSD limit ation. It should insert them strategically to benefit load-balancing along the l ongest part of the path.
</t> </t>
<t> <t pn="section-7.1.1-4">
PE1 can take into account multiple parameters when inserting ELs, as exam PE1 can take into account multiple parameters when inserting ELs; as exam
ples: ples:
<list style="symbols"> </t>
<t>The ERLD value advertised by transit nodes.</t> <ul spacing="normal" bare="false" empty="false" pn="section-7.1.1-5">
<t>The requirement of load-balancing for a particular label value.</t> <li pn="section-7.1.1-5.1">The ERLD value advertised by transit node
<t>Any service provider preference: favor beginning of the path or end of s.</li>
the path.</t> <li pn="section-7.1.1-5.2">The requirement of load-balancing for a p
</list> articular label value.</li>
</t> <li pn="section-7.1.1-5.3">Any service provider preference: favor be
<t> ginning of the path or end of the path.</li>
In Figure 5, a good strategy may be to use the following stack &lt;Adj_P1 </ul>
P2, Adj_set_P2P3, ELI1, EL1, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_P6PE2, ELI2, EL2, <t pn="section-7.1.1-6">
VPN_label&gt;. In <xref target="fig_ex1" format="default" sectionFormat="of" derivedCont
The original stack requests P2 to forward based on a L3 adjacency set that will ent="Figure 5"/>, a good strategy may be to use the following stack &lt;Adj_P1P2
require load-balancing. Therefore it is important to ensure that P2 can load-bal , Adj_set_P2P3, ELI1, EL1, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_P6PE2, ELI2, EL2, V
ance correctly. PN_label&gt;.
The original stack requests P2 to forward based on an L3 adjacency-set that will
require load-balancing. Therefore, it is important to ensure that P2 can load-b
alance correctly.
As P2 has a limited ERLD of 3, an ELI/EL must be inserted just after the label t hat P2 will use to forward. As P2 has a limited ERLD of 3, an ELI/EL must be inserted just after the label t hat P2 will use to forward.
On the path to PE2, P3 has also a limited ERLD, but P3 will forward based on a r egular adjacency segment that may not require load-balancing. On the path to PE2, P3 has also a limited ERLD, but P3 will forward based on a r egular adjacency segment that may not require load-balancing.
Therefore it does not seem important to ensure that P3 can do load-balancing des Therefore, it does not seem important to ensure that P3 can do load-balancing de
pite its limited ERLD. spite its limited ERLD.
The next nodes along the forwarding path have a high ERLD that does not cause an The next nodes along the forwarding path have a high ERLD that does not cause
y issue, except P6. Moreover, P6 is using some LAGs to PE2 and so is expected to any issue, except P6. Moreover, P6 is using some LAGs to PE2 and so is
load-balance. expected to load-balance.
It becomes important to insert a new ELI/EL just after the P6 forwarding label. It becomes important to insert a new ELI/EL just after the P6 forwarding label.
</t> </t>
<t> <t pn="section-7.1.1-7">
In the case above, the ingress node had a sufficient MSD to ensure end-to In the case above, the ingress node was able to support a sufficient MSD
-end load-balancing taking into the path attributes. to ensure
end-to-end load-balancing while taking into account the path attributes.
However, there might be cases where the ingress node may not have the necessary label imposition capacity. However, there might be cases where the ingress node may not have the necessary label imposition capacity.
</t> </t>
</section> </section>
<section anchor="ex2" title="Example 2 where the ingress node does not ha <section anchor="ex2" numbered="true" toc="include" removeInRFC="false"
ve a sufficient MSD"> pn="section-7.1.2">
<name slugifiedName="name-example-2-the-ingress-node-">Example 2: The
<figure title="Figure 6"> Ingress Node Does Not Have a Sufficient MSD</name>
<artwork> <figure anchor="fig_ex2" align="left" suppress-title="false" pn="figur
e-6">
<name slugifiedName="name-msd-considerations">MSD Considerations</na
me>
<artwork name="" type="" align="left" alt="" pn="section-7.1.2-1.1">
ECMP LAG ECMP ECMP ECMP LAG ECMP ECMP
PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- P7 --- P8 --- PE2 PE1 --- P1 --- P2 --- P3 --- P4 --- P5 --- P6 --- P7 --- P8 --- PE2
</artwork>
</artwork> </figure>
</figure> <t pn="section-7.1.2-2">
<t> In <xref target="fig_ex2" format="default" sectionFormat="of" derivedCont
In Figure 6, PE1 wants to forward MPLS VPN traffic over an explicit path ent="Figure 6"/>, PE1 wants to forward MPLS VPN traffic over an explicit path to
to PE2 resulting in the following label stack to be pushed onto the IP header: & PE2 resulting in the following label stack to be pushed onto the IP header: &lt
lt;Adj_P1P2, Adj_set_P2P3, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_set_P6P7, Adj_P7P8; ;Adj_P1P2, Adj_set_P2P3, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_set_P6P7, Adj_P7P8; A
Adj_set_P8PE2, VPN_label&gt;. dj_set_P8PE2, VPN_label&gt;.
PE1 is limited to push a maximum of 11 labels. P2, P3 and P6 have an ERLD of 3 w PE1 is limited to push a maximum of 11 labels. P2, P3, and P6 have an ERLD of 3
hile others have an ERLD of 15. while others have an ERLD of 15.
</t> </t>
<t> <t pn="section-7.1.2-3">
Using a similar strategy as the previous case may lead to a dilemma, as P E1 can only push a single ELI/EL while we may need a minimum of three to load-ba lance the end-to-end path. Using a similar strategy as the previous case may lead to a dilemma, as P E1 can only push a single ELI/EL while we may need a minimum of three to load-ba lance the end-to-end path.
An optimized stack that would enable end-to-end load-balancing may be: &lt;Adj_P 1P2, Adj_set_P2P3, ELI1, EL1, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_set_P6P7, ELI2, EL2, Adj_P7P8, Adj_set_P8PE2, ELI3, EL3, VPN_label&gt;. An optimized stack that would enable end-to-end load-balancing may be: &lt;Adj_P 1P2, Adj_set_P2P3, ELI1, EL1, Adj_P3P4, Adj_P4P5, Adj_P5P6, Adj_set_P6P7, ELI2, EL2, Adj_P7P8, Adj_set_P8PE2, ELI3, EL3, VPN_label&gt;.
</t> </t>
<t> <t pn="section-7.1.2-4">
A decision needs to be taken to favor some part of the path for load-bala ncing considering that load-balancing may not work on the other parts. A decision needs to be taken to favor some part of the path for load-bala ncing considering that load-balancing may not work on the other parts.
A service provider may decide to place the ELI/EL after the P6 forwarding label A service provider may decide to place the ELI/EL after the P6 forwarding
as it will allow P4 and P6 to load-balance. Placing the ELI/EL at bottom of the label as it will allow P4 and P6 to load-balance. Placing the ELI/EL at the bott
stack is also a possibility enabling load-balancing for P4 and P8. om of the stack is also a possibility enabling load-balancing for P4 and P8.
</t> </t>
</section> </section>
</section> </section>
<section anchor="el_placement" title="Considerations for the placement of <section anchor="el_placement" numbered="true" toc="include" removeInRFC="
entropy labels"> false" pn="section-7.2">
<t> <name slugifiedName="name-considerations-for-the-plac">Considerations fo
The sample cases described in the previous section showed that ELI/EL pla r the Placement of Entropy Labels</name>
cement when the maximum number of labels to be pushed is limited is not an easy <t pn="section-7.2-1">
decision and multiple criteria may be taken into account. The sample cases described in the previous section showed that ELI/EL pla
</t> cement when the maximum number of labels to be pushed is limited is not an easy
<t> decision, and multiple criteria may be taken into account.
This section describes some considerations that an implementation MAY tak </t>
e into account when placing ELI/ELs. This list of criteria is not considered exh <t pn="section-7.2-2">
austive and an implementation MAY take into account additional criteria or tie-b This section describes some considerations that an implementation <bcp14>
reakers that are not documented here. MAY</bcp14> take into account when placing ELI/ELs. This list of criteria is not
considered exhaustive and an implementation <bcp14>MAY</bcp14> take into accoun
t additional criteria or tiebreakers that are not documented here.
As the insertion of ELI/ELs is performed by the ingress node, having ingr ess nodes that do not use the same criteria does not cause an interoperability i ssue. However, from a network design and operation perspective, it is better to have all ingress routers using the same criteria. As the insertion of ELI/ELs is performed by the ingress node, having ingr ess nodes that do not use the same criteria does not cause an interoperability i ssue. However, from a network design and operation perspective, it is better to have all ingress routers using the same criteria.
</t> </t>
<t> <t pn="section-7.2-3">
An implementation SHOULD try to maximize the possibility of load-balancin An implementation <bcp14>SHOULD</bcp14> try to maximize the possibility o
g along the path by inserting an ELI/EL where multiple equal cost paths are avai f load-balancing along the path by inserting an ELI/EL where multiple equal-cost
lable and minimize the number of ELI/ELs that need to be inserted. paths are available and minimize the number of ELI/ELs that need to be inserted
In case of a trade-off, an implementation SHOULD provide flexibility to the .
operator to select the criteria to be considered when placing ELI/ELs or specify In case of a trade-off, an implementation <bcp14>SHOULD</bcp14> provide flex
a sub-objective for optimization. ibility to the operator to select the criteria to be considered when placing ELI
</t> /ELs or specify a subobjective for optimization.
</t>
<figure title="Figure 7"> <figure anchor="fig_consid_sample" align="left" suppress-title="false" p
<artwork> n="figure-7">
<name slugifiedName="name-msd-trade-offs">MSD Trade-Offs</name>
<artwork name="" type="" align="left" alt="" pn="section-7.2-4.1">
2 2 2 2
PE1 -- P1 -- P2 --P3 --- P4 --- P5 -- ... -- P8 -- P9 -- PE2 PE1 -- P1 -- P2 --P3 --- P4 --- P5 -- ... -- P8 -- P9 -- PE2
| | | |
P3'--- P4'--- P5' P3'--- P4'--- P5'
</artwork> </artwork>
</figure> </figure>
<t> <t pn="section-7.2-5"><xref target="fig_consid_sample" format="default"
Figure 7 will be used as reference in the following subsections. All metr sectionFormat="of" derivedContent="Figure 7"/> will be used as reference in the
ics are equal to 1, except P3-P4 and P4-P5 which have a metric 2. following subsections. All
We consider the MSD of nodes to be the full limit of label imposition (in metrics are equal to 1 except P3-P4 and P4-P5, which have a metric 2.
cluding service labels, entropy labels and transport labels). We consider the MSD of nodes to be the full limit of label imposition
</t> (including service labels, entropy labels, and transport labels).
<section anchor="erld" title="ERLD value"> </t>
<t> <section anchor="erld" numbered="true" toc="include" removeInRFC="false"
As mentioned in <xref target="overview"/>, the ERLD value is an i pn="section-7.2.1">
mportant parameter to consider when inserting an ELI/EL. If an ELI/EL does not f <name slugifiedName="name-erld-value">ERLD Value</name>
all within the ERLD of a node on the path, the node will not be able to load-bal <t pn="section-7.2.1-1">
ance the traffic efficiently. As mentioned in <xref target="overview" format="default" sectionF
</t> ormat="of" derivedContent="Section 7.1"/>, the ERLD value is an important parame
<t> ter to consider when inserting an ELI/EL. If an ELI/EL does not fall within the
The ERLD value can be advertised via protocols and those extensio ERLD of a node on the path, the node will not be able to load-balance the traffi
ns are described in separate documents (for instance, <xref target="I-D.ietf-isi c efficiently.
s-mpls-elc"/> and <xref target="I-D.ietf-ospf-mpls-elc"/>). </t>
</t> <t pn="section-7.2.1-2">
<t> The ERLD value can be advertised via protocols, and those extensi
ons are described in separate documents (for instance, <xref target="I-D.ietf-is
is-mpls-elc" format="default" sectionFormat="of" derivedContent="ISIS-ELC"/> and
<xref target="I-D.ietf-ospf-mpls-elc" format="default" sectionFormat="of" deriv
edContent="OSPF-ELC"/>).
</t>
<t pn="section-7.2.1-3">
Let's consider a path from PE1 to PE2 using the following stack p ushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2, Service_label&gt;. Let's consider a path from PE1 to PE2 using the following stack p ushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2, Service_label&gt;.
</t> </t>
<t> <t pn="section-7.2.1-4">
Using the ERLD as an input parameter can help to minimize the num ber of required ELI/EL pairs to be inserted. Using the ERLD as an input parameter can help to minimize the num ber of required ELI/EL pairs to be inserted.
An ERLD value must be retrieved for each SPRING label in the label stack . An ERLD value must be retrieved for each SPRING label in the label stack .
</t> </t>
<t> <t pn="section-7.2.1-5">
For a label bound to an adjacency segment, the ERLD is the ERLD o For a label bound to an adjacency segment, the ERLD is the ERLD o
f the node that has advertised the adjacency segment. In the example above, the f the node that has advertised the adjacency segment. In the example above, the
ERLD associated with Adj_P1P2 would be the ERLD of router P1 as P1 will perform ERLD associated with Adj_P1P2 would be the ERLD of router P1, as P1 will perform
the forwarding based on the Adj_P1P2 label. the forwarding based on the Adj_P1P2 label.
</t> </t>
<t> <t pn="section-7.2.1-6">
For a label bound to a node segment, multiple strategies MAY be i For a label bound to a node segment, multiple strategies <bcp14>M
mplemented. An implementation MAY try to evaluate the minimum ERLD value along t AY</bcp14> be implemented. An implementation <bcp14>MAY</bcp14> try to evaluate
he node segment path. the minimum ERLD value along the node segment path.
If an implementation cannot find the minimum ERLD along the path of the segment If an implementation cannot find the minimum ERLD along the path of the
or does not support the computation of the minimum ERLD, it SHOULD instead use t segment or does not support the computation of the minimum ERLD, it <bcp14>SHOUL
he ERLD of the tail-end node. Using the ERLD of the tail-end of the node segment D</bcp14>
mimics the behavior of <xref target="RFC6790"/> where the ingress takes only ca instead use the ERLD of the tail-end node. Using the ERLD of the tail end of the
re of the egress of the LSP. node segment mimics the behavior of <xref target="RFC6790" format="default" sec
tionFormat="of" derivedContent="RFC6790"/> where the ingress takes only care of
the egress of the LSP.
In the example above, if the implementation supports computation of minimum ERLD along the path, the ERLD associated with label Node_P9 would be the minimum ERL D between nodes {P2,P3,P4 ..., P8}. In the example above, if the implementation supports computation of minimum ERLD along the path, the ERLD associated with label Node_P9 would be the minimum ERL D between nodes {P2,P3,P4 ..., P8}.
If the implementation does not support the computation of minimum ERLD, it will If the implementation does not support the computation of minimum ERLD, it
consider the ERLD of P9 (tail-end node of Node_P9 SID). While providing the more will consider the ERLD of P9 (tail-end node of Node_P9 SID). While providing
optimal ELI/EL placement, evaluating the minimum ERLD increases the complexity the more optimal ELI/EL placement, evaluating the minimum ERLD increases the
of ELI/EL insertion. As the path to the Node-SID may change over time, a recompu complexity of ELI/EL insertion. As the path to the Node SID may change over time
tation of the minimum ERLD is required for each topology change. This recomputat , a recomputation of the minimum ERLD is required for each topology change. This
ion may require the positions of the ELI/ELs to change. recomputation may require the positions of the ELI/ELs to change.
</t> </t>
<t> <t pn="section-7.2.1-7">
For a label bound to a binding segment, if the binding segment describes a path, For a label bound to a Binding Segment, if the Binding Segment describes a
an implementation MAY also try to evaluate the minimum ERLD along this path. If path, an implementation <bcp14>MAY</bcp14> also try to evaluate the minimum ERLD
the implementation cannot find the minimum ERLD along the path of the segment o along this
r does not support this evaluation, it SHOULD instead use the ERLD of the node a path. If the implementation cannot find the minimum ERLD along the path of the
dvertising the Binding-SID. segment or does not support this evaluation, it <bcp14>SHOULD</bcp14> instead us
As for the node segment, evaluating the minimum ERLD adds complexity in the ELI/ e the ERLD of
EL insertion process. the node advertising the Binding SID. As for the node segment, evaluating the
</t> minimum ERLD adds complexity in the ELI/EL insertion process.
</section> </t>
<section anchor="sid-type" title="Segment type"> </section>
<t> <section anchor="sid-type" numbered="true" toc="include" removeInRFC="fa
Depending on the type of segment a particular label is bound to, lse" pn="section-7.2.2">
an implementation can deduce that this particular label will be subject to load- <name slugifiedName="name-segment-type">Segment Type</name>
balancing on the path. <t pn="section-7.2.2-1">
</t> Depending on the type of segment a particular label is bound
<section anchor="node-sid" title="Node-SID"> to, an implementation can deduce that this particular label
<t> will be subject to load-balancing on the path.
An MPLS label bound to a Node-SID represents a path that </t>
may cross multiple hops. <section anchor="node-sid" numbered="true" toc="exclude" removeInRFC="
Load-balancing may be needed on the node starting this path but also on any node false" pn="section-7.2.2.1">
along the path. <name slugifiedName="name-node-sid">Node SID</name>
</t> <t pn="section-7.2.2.1-1">
<t> An MPLS label bound to a Node SID represents a path
In Figure 7, let's consider a path from PE1 to PE2 using that may cross multiple hops. Load-balancing may be
the following stack pushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2, Service_lab needed on the node starting this path but also on any
el&gt;. node along the path.
</t> </t>
<t> <t pn="section-7.2.2.1-2">
If, for example, PE1 is limited to push 6 labels, it can In <xref target="fig_consid_sample" format="default" sect
add a single ELI/EL within the label stack. ionFormat="of" derivedContent="Figure 7"/>, let's consider a path from PE1 to PE
An operator may want to favor a placement that would allow load-balancing along 2 using the following stack pushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2, Ser
the Node-SID path. vice_label&gt;.
In Figure 7, P3 which is along the Node-SID path requires load-balancing between </t>
two equal-cost paths. <t pn="section-7.2.2.1-3">
</t> If, for example, PE1 is limited to push 6 labels, it
<t> can add a single ELI/EL within the label stack. An
An implementation MAY try to evaluate if load-balancing is really operator may want to favor a placement that would
allow load-balancing along the Node SID path. In
<xref target="fig_consid_sample" format="default" section
Format="of" derivedContent="Figure 7"/>,
P3, which is along the Node SID path,
requires load-balancing between two equal-cost paths.
</t>
<t pn="section-7.2.2.1-4">
An implementation <bcp14>MAY</bcp14> try to evaluate if load-balancing is really
required within a node segment path. This could be done by running required within a node segment path. This could be done by running
an additional SPT (Shortest Path Tree) computation and analysing of the node segment path to an additional SPT (Shortest Path Tree) computation and analyzing of the node segment path to
prevent a node segment that does not really require load-balancing from prevent a node segment that does not really require load-balancing from
being preferred when placing ELI/ELs. Such inspection may be time being preferred when placing ELI/ELs. Such inspection may be time
consuming for implementations and without a 100% guarantee, as a node consuming for implementations and without a 100% guarantee, as a node
segment path may use LAGs that are invisible to the IP segment path may use LAGs that are invisible to the IP
topology. As a simpler approach, an implementation MAY consider that a label topology. As a simpler approach, an implementation <bcp14>MAY</bcp14> consid
bound er that a label bound
to a Node-SID will be subject to load-balancing and requires an to a Node SID will be subject to load-balancing and require an
ELI/EL. ELI/⁠EL.
</t> </t>
</section> </section>
<section anchor="adj-sid1" title="Adjacency-set SID"> <section anchor="adj-sid1" numbered="true" toc="exclude" removeInRFC="
<t> false" pn="section-7.2.2.2">
An adjacency-set is an Adj-SID that refers to a set of ad <name slugifiedName="name-adjacency-set-sid">Adjacency-Set SID</name
jacencies. When an adjacency-set segment is used within a label stack, an implem >
entation can deduce that load-balancing is expected at the node that advertised <t pn="section-7.2.2.2-1">
this adjacency segment. An adjacency-set is an Adj-SID that refers to a set of
An implementation MAY favor the insertion of an ELI/EL af adjacencies. When an adjacency-set segment is used
ter the Adj-SID representing an adjacency-set. within a label stack, an implementation can deduce
</t> that load-balancing is expected at the node that
</section> advertised this adjacency segment. An implementation
<section anchor="adj-sid2" title="Adjacency-SID represent <bcp14>MAY</bcp14> favor the insertion of an ELI/EL
ing a single IP link"> after the Adj-SID representing an adjacency-set.
<t> </t>
</section>
<section anchor="adj-sid2" numbered="true" toc="exclude" removeInRFC="
false" pn="section-7.2.2.3">
<name slugifiedName="name-adjacency-sid-representing-">Adjacency SID
Representing a Single IP Link</name>
<t pn="section-7.2.2.3-1">
When an adjacency segment representing a single IP link i s used within a label stack, an implementation can deduce that load-balancing ma y not be expected at the node that advertised this adjacency segment. When an adjacency segment representing a single IP link i s used within a label stack, an implementation can deduce that load-balancing ma y not be expected at the node that advertised this adjacency segment.
</t> </t>
<t> <t pn="section-7.2.2.3-2">
An implementation MAY NOT place an ELI/EL after a regular An implementation <bcp14>MAY</bcp14> NOT place an ELI/EL
Adj-SID in order to favor the insertion of ELI/ELs following other segments. after a regular Adj-SID in order to favor the insertion of ELI/ELs following oth
</t> er segments.
<t> </t>
Readers should note that an adjacency segment representin <t pn="section-7.2.2.3-3">
g a single IP link may require load-balancing. This is the case when a LAG (L2 b Readers should note that an adjacency segment representin
undle) is implemented between two IP nodes and the L2 bundle SR extensions <xref g a single IP link may require load-balancing. This is the case when a LAG (L2 b
target="I-D.ietf-isis-l2bundles"/> are not implemented. undle) is implemented between two IP nodes and the L2 bundle SR extensions <xref
target="RFC8668" format="default" sectionFormat="of" derivedContent="RFC8668"/>
are not implemented.
In such a case, it could be useful to insert an ELI/EL in a readable position for the LSR advertising the label associated with the adjac ency segment. In such a case, it could be useful to insert an ELI/EL in a readable position for the LSR advertising the label associated with the adjac ency segment.
To communicate the requirement for load-balancing for a p To communicate the requirement for load-balancing for
articular Adjacency-SID to ingress nodes, a user can enforce the use of the L2 b a particular Adjacency SID to ingress nodes, a user can e
undle SR extensions defined in <xref target="I-D.ietf-isis-l2bundles"/> or can d nforce the use of the L2 bundle SR extensions defined in <xref target="RFC8668"
eclare the single adjacency as an adjacency-set. format="default" sectionFormat="of" derivedContent="RFC8668"/> or can declare th
</t> e single adjacency as an adjacency-set.
</section> </t>
<section anchor="adj-sid3" title="Adjacency-SID represent </section>
ing a single link within an L2 bundle"> <section anchor="adj-sid3" numbered="true" toc="exclude" removeInRFC="
<t> false" pn="section-7.2.2.4">
When the L2 bundle SR extensions <xref target="I-D.ietf-i <name slugifiedName="name-adjacency-sid-representing-a">Adjacency SI
sis-l2bundles"/> are used, adjacency segments may be advertised for each member D Representing a Single Link within an L2 Bundle</name>
of the bundle. <t pn="section-7.2.2.4-1">
In this case, an implementation can deduce that load-bala When the L2 bundle SR extensions <xref target="RFC8668" f
ncing is not expected on the LSR advertising this segment and MAY NOT insert an ormat="default" sectionFormat="of" derivedContent="RFC8668"/> are used, adjacenc
ELI/EL after the corresponding label. y segments may be advertised for each member of the bundle.
</t> In this case, an implementation can deduce that load-bala
</section> ncing is not expected on the LSR advertising this segment and <bcp14>MAY</bcp14>
<section anchor="adj-sid4" title="Adjacency-SID represent NOT insert an ELI/EL after the corresponding label.
ing an L2 bundle"> </t>
<t> </section>
When the L2 bundle SR extensions <xref target="I-D.ietf-i <section anchor="adj-sid4" numbered="true" toc="exclude" removeInRFC="
sis-l2bundles"/> are used, an adjacency segment may be advertised to represent t false" pn="section-7.2.2.5">
he bundle. <name slugifiedName="name-adjacency-sid-representing-an">Adjacency S
In this case, an implementation can deduce that load-balancing is expected on th ID Representing an L2 Bundle</name>
e LSR advertising this segment and MAY insert an ELI/EL after the corresponding <t pn="section-7.2.2.5-1">
label. When the L2 bundle SR extensions <xref target="RFC8668" f
</t> ormat="default" sectionFormat="of" derivedContent="RFC8668"/> are used, an adjac
</section> ency segment may be advertised to represent the bundle.
In this case, an implementation can deduce that load-balancing is expected on th
</section> e LSR advertising this segment and <bcp14>MAY</bcp14> insert an ELI/EL after the
<section title="Maximizing number of LSRs that will load-balance" corresponding label.
> </t>
<t> </section>
When placing ELI/ELs, an implementation MAY optimize the </section>
number of LSRs that both need to load-balance (i.e., have <section numbered="true" toc="include" removeInRFC="false" pn="section-7
ECMP paths) and that will be able to perform load-balancing (i.e., .2.3">
the EL label is within their ERLD). <name slugifiedName="name-maximizing-number-of-lsrs-t">Maximizing Numb
</t> er of LSRs That Will Load-Balance</name>
<t> <t pn="section-7.2.3-1">
Let's consider a path from PE1 to PE2 using the following stack p When placing ELI/ELs, an implementation <bcp14>MAY</bcp14>
ushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2, Service_label&gt;. optimize the number of LSRs that both need to load-balance
All routers have an ERLD of 10, except P1 and P2 which have an ERLD of 4. PE1 is (i.e., have ECMPs) and that will be able to perform
able to push 6 labels, so only a single ELI/EL can be added. load-balancing (i.e., the EL is within their ERLD).
</t> </t>
<t> <t pn="section-7.2.3-2">
In the example above, adding an ELI/EL after Adj_P1P2 will only a Let's consider a path from PE1 to PE2 using the following
llow load-balancing at P1 while inserting it after Adj_PE2P9, will allow load-ba stack pushed by PE1: &lt;Adj_P1P2, Node_P9, Adj_P9PE2,
lancing at P2,P3 ... P9 and maximizing the number of LSRs that can perform load- Service_label&gt;. All routers have an ERLD of 10 except P1
balancing. and P2, which have an ERLD of 4. PE1 is able to push 6 labels,
</t> so only a single ELI/EL can be added.
</section> </t>
<section title="Preference for a part of the path"> <t pn="section-7.2.3-3">
<t> In the example above, adding an ELI/EL after Adj_P1P2 will
An implementation MAY allow the user to favor a part of the end-t only allow load-balancing at P1, while inserting it after
o-end path when the number of ELI/ELs that can be pushed is not enough to cover Adj_PE2P9 will allow load-balancing at P2, P3 ... P9 and
the entire path. maximize the number of LSRs that can perform load-balancing.
As an example, a service provider may want to favor load-balancing at the beginn </t>
ing of the path or at the end of path, so the implementation favors putting the </section>
ELI/ELs near the top or near of the bottom of the stack. <section numbered="true" toc="include" removeInRFC="false" pn="section-7
</t> .2.4">
</section> <name slugifiedName="name-preference-for-a-part-of-th">Preference for
<section title="Combining criteria"> a Part of the Path</name>
<t> <t pn="section-7.2.4-1">
An implementation MAY combine multiple criteria to determine the An implementation <bcp14>MAY</bcp14> allow the user to favor a pa
best ELI/ELs placement. However, combining too many criteria could lead to imple rt of the end-to-end path when the number of ELI/ELs that can be pushed is not e
mentation complexity and high resource consumption. nough to cover the entire path.
Each time the network topology changes, a new As an example, a service provider may want to favor load-balancing at the
evaluation of the ELI/EL placement will be necessary for each beginning of the path or at the end of the path, so the implementation favors
impacted LSPs. putting the ELI/ELs near the top or the bottom of the stack.
</t> </t>
</section> </section>
</section> <section numbered="true" toc="include" removeInRFC="false" pn="section-7
</section> .2.5">
<section anchor="algo-example" title="A simple example algorithm" toc="default" <name slugifiedName="name-combining-criteria">Combining Criteria</name
> >
<t> <t pn="section-7.2.5-1">
A simple implementation might take into account the ERLD when placing ELI/EL wh An implementation <bcp14>MAY</bcp14> combine multiple criteria to
ile trying to minimize the number of ELI/ELs inserted and trying to maximize the determine
number of LSRs that can load-balance. the best ELI/ELs placement. However, combining too many
</t> criteria could lead to implementation complexity and high
<t> resource consumption. Each time the network topology changes,
a new evaluation of the ELI/EL placement will be necessary for
each impacted LSP.
</t>
</section>
</section>
</section>
<section anchor="algo-example" toc="include" numbered="true" removeInRFC="fa
lse" pn="section-8">
<name slugifiedName="name-a-simple-example-algorithm">A Simple Example Alg
orithm</name>
<t pn="section-8-1">
A simple implementation might take into account the ERLD when placing ELI/EL
while trying to minimize the number of ELI/ELs inserted and trying to
maximize the number of LSRs that can load-balance.
</t>
<t pn="section-8-2">
The example algorithm is based on the following considerations: The example algorithm is based on the following considerations:
<list style="symbols"> </t>
<t>An LSR that can insert a limited number of &lt;ELI, EL&gt; pairs should inse <ul spacing="normal" bare="false" empty="false" pn="section-8-3">
rt such pairs deeper in the stack.</t> <li pn="section-8-3.1">An LSR that can insert a limited number of &lt;EL
<t>An LSR should try to insert &lt;ELI, EL&gt; pairs at positions to maximize t I, EL&gt; pairs should insert such pairs deeper in the stack.</li>
he number of transit LSRs for which the EL occurs within the ERLD of those LSRs. <li pn="section-8-3.2">An LSR should try to insert &lt;ELI, EL&gt; pairs
</t> at positions to maximize the number of transit LSRs for which the EL occurs wit
<t>An LSR should try to insert the minimum number of such pairs while trying to hin the ERLD of those LSRs.</li>
satisfy the above criteria.</t> <li pn="section-8-3.3">An LSR should try to insert the minimum number of
</list> such pairs while trying to satisfy the above criteria.</li>
</t> </ul>
<t> <t pn="section-8-4">
The pseudocode of the example algorithm is shown below. The pseudocode of the example algorithm is shown below.
</t> </t>
<figure title="Figure 8: Example algorithm to insert &lt;ELI, EL&gt; pairs in a <figure align="left" suppress-title="false" pn="figure-8">
label stack"> <name slugifiedName="name-example-algorithm-to-insert">Example Algorithm
<artwork> to Insert &lt;ELI, EL&gt; Pairs in a Label Stack</name>
Initialize the current EL insertion point to the <sourcecode type="pseudocode" markers="false" pn="section-8-5.1">
bottom-most label in the stack that is EL-capable Initialize the current EL insertion point to the
while (local-node can push more &lt;ELI,EL&gt; pairs OR bottom-most label in the stack that is EL-capable
insertion point is not above label stack) { while (local-node can push more &lt;ELI,EL&gt; pairs OR
insert an &lt;ELI,EL&gt; pair below current insertion point insertion point is not above label stack) {
move new insertion point up from current insertion point until insert an &lt;ELI,EL&gt; pair below current insertion point
((last inserted EL is below the ERLD) AND (ERLD > 2) move new insertion point up from current insertion point until
AND ((last inserted EL is below the ERLD) AND (ERLD &gt; 2)
(new insertion point is EL-capable)) AND
set current insertion point to new insertion point (new insertion point is EL-capable))
} set current insertion point to new insertion point
</artwork> }
</figure> </sourcecode>
<t> </figure>
When this algorithm is applied to the example described in <xref target="usecas <t pn="section-8-6">
e"/>, When this algorithm is applied to the example described in <xref target="usecas
it will result in ELs being inserted in two positions, one after the e" format="default" sectionFormat="of" derivedContent="Section 3"/>,
it will result in ELs being inserted in two positions; one after the
label L_N-D and another after L_N-P3. Thus, the resulting label stack label L_N-D and another after L_N-P3. Thus, the resulting label stack
would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI, EL&gt; would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI, EL&gt;.
</t> </t>
</section>
</section> <section anchor="deployment" numbered="true" toc="include" removeInRFC="fals
e" pn="section-9">
<section anchor="deployment" title="Deployment Considerations"> <name slugifiedName="name-deployment-considerations">Deployment Considerat
<t> ions</name>
As long as LSR node dataplane capabilities are limited (number of labels that c <t pn="section-9-1">
an be pushed, or number of labels that can be inspected), hop-by-hop load-balanc As long as LSR node data-plane capabilities are limited (number of labels that
ing of SPRING encapsulated flows will require trade-offs. can be pushed or number of labels that can be inspected), hop-by-hop load-balanc
</t> ing of SPRING-encapsulated flows will require trade-offs.
<t> </t>
The entropy label is still a good and usable solution as it allows load-balanci <t pn="section-9-2">
ng without having to perform deep packet inspection on each LSR: it does not see The entropy label is still a good and usable solution as it allows load-balanci
m reasonable to have an LSR inspecting UDP ports within a GRE tunnel carried ove ng without having to perform deep packet inspection on each LSR: It does not see
r a 15 label SPRING tunnel. m reasonable to have an LSR inspecting UDP ports within a GRE tunnel carried ove
</t> r a 15-label SPRING tunnel.
<t> </t>
Due to the limited capacity of reading a deep stack of MPLS labels, multiple EL <t pn="section-9-3">
I/ELs may be required within the stack which directly impacts the capacity of th Due to the limited capacity of reading a deep stack of MPLS labels, multiple EL
e head-end to push a deep stack: each ELI/EL inserted requires two additional la I/ELs may be required within the stack, which directly impacts the capacity of t
bels to be pushed. he head-end to push a deep stack: each ELI/EL inserted requires two additional l
</t> abels to be pushed.
<t> </t>
Placement strategies of ELI/ELs are required to find the best trade-off. Multip <t pn="section-9-4">
le criteria could be taken into account and some level of customization (by the Placement strategies of ELI/ELs are required to find the best trade-off. Multip
user) is required to accommodate different deployments. le criteria could be taken into account, and some level of customization (by the
user) is required to accommodate different deployments.
Since analyzing the path of each destination to determine the best ELI/EL place ment may be time consuming for the control plane, we encourage implementations t o find the best trade-off between simplicity, resource consumption, and load-bal ancing efficiency. Since analyzing the path of each destination to determine the best ELI/EL place ment may be time consuming for the control plane, we encourage implementations t o find the best trade-off between simplicity, resource consumption, and load-bal ancing efficiency.
</t> </t>
<t> <t pn="section-9-5">
In the future, hardware and software capacity may increase dataplane capabiliti In the future, hardware and software capacity may increase data-plane capabilit
es and may remove some of these limitations, increasing load-balancing capabilit ies and may remove some of these limitations, increasing load-balancing capabili
y using entropy labels. ty using entropy labels.
</t> </t>
</section> </section>
<section anchor="other-options" title="Options considered"> <section anchor="other-options" numbered="true" toc="include" removeInRFC="f
<t>Different options that were considered to arrive at the recommended alse" pn="section-10">
<name slugifiedName="name-options-considered">Options Considered</name>
<t pn="section-10-1">Different options that were considered to arrive at t
he recommended
solution are documented in this section. solution are documented in this section.
</t> </t>
<t> <t pn="section-10-2">
These options are detailed here only for historical purposes. These options are detailed here only for historical purposes.
</t> </t>
<section title="Single EL at the bottom of the stack"> <section numbered="true" toc="include" removeInRFC="false" pn="section-10.
<t> 1">
<name slugifiedName="name-single-el-at-the-bottom-of-">Single EL at the
Bottom of the Stack</name>
<t pn="section-10.1-1">
In this option, a single EL is used for the entire label stack. The In this option, a single EL is used for the entire label stack. The
source LSR S encodes the entropy label at the bottom of the source LSR S encodes the entropy label at the bottom of the
label stack. In the example described in <xref target="usecase"/>, it will r esult label stack. In the example described in <xref target="usecase" format="defa ult" sectionFormat="of" derivedContent="Section 3"/>, it will result
in the label stack at LSR S to look like &lt;L_N-P3, L_A-L1, L_N-D, ELI, in the label stack at LSR S to look like &lt;L_N-P3, L_A-L1, L_N-D, ELI,
EL&gt; &lt;remaining packet header&gt;. Note that the notation in <xref targ et="RFC6790"/> EL&gt; &lt;remaining packet header&gt;. Note that the notation in <xref targ et="RFC6790" format="default" sectionFormat="of" derivedContent="RFC6790"/>
is used to describe the label stack. An issue with this approach is is used to describe the label stack. An issue with this approach is
that as the label stack grows due an increase in the number of SIDs, that as the label stack grows due an increase in the number of SIDs,
the EL goes correspondingly deeper in the label stack. Hence, transit the EL goes correspondingly deeper in the label stack. Hence, transit
LSRs have to access a larger number of bytes in the packet header LSRs have to access a larger number of bytes in the packet header
when making forwarding decisions. In the example described in when making forwarding decisions. In the example described in
<xref target="usecase"/>, if we consider that the LSR P1 has an ERLD of 3, P1 would <xref target="usecase" format="default" sectionFormat="of" derivedContent="Se ction 3"/>, if we consider that the LSR P1 has an ERLD of 3, P1 would
load-balance traffic poorly on the load-balance traffic poorly on the
parallel links L3 and L4 since the EL is below the ERLD of P1. parallel links L3 and L4 since the EL is below the ERLD of P1.
A load-balanced network design using this approach A load-balanced network design using this approach
must ensure that all intermediate LSRs have the capability to must ensure that all intermediate LSRs have the capability to
read the maximum label stack depth as required for the read the maximum label stack depth as required for the
application that uses source routed stacking. application that uses source-routed stacking.
</t> </t>
<t> <t pn="section-10.1-2">
This option was rejected since there exist a number of hardware This option was rejected since there exist a number of hardware
implementations which have a low maximum readable label depth. implementations that have a low maximum readable label depth.
Choosing this option can lead to a loss of load-balancing using EL in Choosing this option can lead to a loss of load-balancing using EL in
a significant part of the network when that is a critical requirement a significant part of the network when that is a critical requirement
in a service-provider network. in a service-provider network.
</t> </t>
</section> </section>
<section title="An EL per segment in the stack"> <section numbered="true" toc="include" removeInRFC="false" pn="section-10.
<t> 2">
In this option, each segment/label in the stack can be given its own EL. <name slugifiedName="name-an-el-per-segment-in-the-st">An EL per Segment
When in the Stack</name>
load-balancing is required to direct traffic on a segment, the <t pn="section-10.2-1">
source LSR pushes an &lt;ELI, EL&gt; before pushing the label associated to t In this option, each segment/label in the stack can be given its own
his segment . In the EL. When load-balancing is required to direct traffic on a segment,
example described in <xref target="usecase"/>, the source LSR S encoded label the source LSR pushes an &lt;ELI, EL&gt; before pushing the label
stack associated to this segment. In the example described in <xref target="us
would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI, EL&gt; where all the ELs ecase" format="default" sectionFormat="of" derivedContent="Section 3"/>, the sou
can be the same. Accessing the EL at an intermediate LSR is rce label stack that is LSR S encoded would
independent of the depth of the label stack and hence independent of be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI, EL&gt;, where all the ELs
the specific application that uses source routed tunnels with label can be the same. Accessing the EL at an intermediate LSR is
stacking. A drawback is that the depth of the label independent of the depth of the label stack and hence, independent of
stack grows significantly, almost 3 times as the number of labels in the specific application that uses source-routed tunnels with label
the label stack. The network design should ensure that source LSRs stacking. A drawback is that the depth of the label stack grows
have the capability to push such a deep label stack. Also, significantly, almost 3 times as the number of labels in the label
the bandwidth overhead and potential MTU issues of deep label stacks stack. The network design should ensure that source LSRs have the
should be considered in the network design. capability to push such a deep label stack. Also, the bandwidth
</t> overhead and potential MTU issues of deep label stacks should be
<t> considered in the network design.
</t>
<t pn="section-10.2-2">
This option was rejected due to the existence of hardware This option was rejected due to the existence of hardware
implementations that can push a limited number of labels on the label implementations that can push a limited number of labels on the label
stack. Choosing this option would result in a hardware requirement stack. Choosing this option would result in a hardware requirement
to push two additional labels per tunnel label. Hence it would to push two additional labels per tunnel label. Hence, it would
restrict the number of tunnels that can be stacked in a LSP and hence restrict the number of tunnels that can be stacked in an LSP and hence,
constrain the types of LSPs that can be created. This was considered constrain the types of LSPs that can be created. This was considered
unacceptable. unacceptable.
</t> </t>
</section> </section>
<section title="A re-usable EL for a stack of tunnels"> <section numbered="true" toc="include" removeInRFC="false" pn="section-10.
<t> 3">
In this option an LSR that terminates a tunnel re-uses the EL of the <name slugifiedName="name-a-reusable-el-for-a-stack-o">A Reusable EL for
terminated tunnel for the next inner tunnel. It does this by storing a Stack of Tunnels</name>
the EL from the outer tunnel when that tunnel is terminated and re- <t pn="section-10.3-1">
inserting it below the next inner tunnel label during the label swap In this option, an LSR that terminates a tunnel reuses the EL of the
operation. The LSR that stacks tunnels should insert an EL below the terminated tunnel for the next inner tunnel. It does this by storing
outermost tunnel. It should not insert ELs for any inner tunnels. the EL from the outer tunnel when that tunnel is terminated and
Also, the penultimate hop LSR of a segment must not pop the ELI and reinserting it below the next inner tunnel label during the label-swap
EL even though they are exposed as the top labels since the operation. The LSR that stacks tunnels should insert an EL below the
terminating LSR of that segment would re-use the EL for the next outermost tunnel. It should not insert ELs for any inner tunnels.
segment. Also, the penultimate hop LSR of a segment must not pop the ELI and EL
</t> even though they are exposed as the top labels since the terminating
<t> LSR of that segment would reuse the EL for the next segment.
In <xref target="usecase"/> above, the source LSR S encoded label stack w </t>
ould be <t pn="section-10.3-2">
&lt;L_N-P3, ELI, EL, L_A-L1, L_N-D&gt;. At P1, the outgoing label stack In <xref target="usecase" format="default" sectionFormat="of" derivedCont
would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D&gt; after it has load-balanced ent="Section 3"/>, the source label stack that is LSR S
to one of the links L3 or L4. At P3 the outgoing label stack would encoded would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-⁠D&gt;. At P1, the
be &lt;L_N-D, ELI, EL&gt;. At P2, the outgoing label stack would be &lt;L_N- outgoing label stack would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D&gt;
D, after it has load-balanced to one of the links L3 or L4. At P3, the
ELI, EL&gt; and it would load-balance to one of the nexthop LSRs P4 or outgoing label stack would be &lt;L_N-D, ELI, EL&gt;. At P2, the
P5. Accessing the EL at an intermediate LSR (e.g., P1) is outgoing label stack would be &lt;L_N-D, ELI, EL&gt; and it would
independent of the depth of the label stack and hence independent of load-balance to one of the next-hop LSRs P4 or P5. Accessing the EL at
the specific use-case to which the label stack is applied. an intermediate LSR (e.g., P1) is independent of the depth of the
</t> label stack and hence, independent of the specific use case to which
<t> the label stack is applied.
This option was rejected due to the significant change in label swap </t>
<t pn="section-10.3-3">
This option was rejected due to the significant change in label-swap
operations that would be required for existing hardware. operations that would be required for existing hardware.
</t> </t>
</section> </section>
<section title="EL at top of stack"> <section numbered="true" toc="include" removeInRFC="false" pn="section-10.
<t> 4">
A slight variant of the re-usable EL option is to keep the EL at the <name slugifiedName="name-el-at-top-of-stack">EL at Top of Stack</name>
<t pn="section-10.4-1">
A slight variant of the reusable EL option is to keep the EL at the
top of the stack rather than below the tunnel label. In this case, top of the stack rather than below the tunnel label. In this case,
each LSR that is not terminating a segment should continue to keep each LSR that is not terminating a segment should continue to keep
the received EL at the top of the stack when forwarding the packet the received EL at the top of the stack when forwarding the packet
along the segment. An LSR that terminates a segment should use the along the segment. An LSR that terminates a segment should use the
EL from the terminated segment at the top of the stack when EL from the terminated segment at the top of the stack when
forwarding onto the next segment. forwarding onto the next segment.
</t> </t>
<t> <t pn="section-10.4-2">
This option was rejected due to the significant change in label swap This option was rejected due to the significant change in label swap
operations that would be required for existing hardware. operations that would be required for existing hardware.
</t> </t>
</section> </section>
<section title="ELs at readable label stack depths"> <section numbered="true" toc="include" removeInRFC="false" pn="section-10.
<t> 5">
In this option the source LSR inserts ELs for tunnels in the label <name slugifiedName="name-els-at-readable-label-stack">ELs at Readable L
stack at depths such that each LSR along the path that must load abel Stack Depths</name>
balance is able to access at least one EL. Note that the source LSR <t pn="section-10.5-1">
may have to insert multiple ELs in the label stack at different In this option, the source LSR inserts ELs for tunnels in the label
depths for this to work since intermediate LSRs may have differing stack at depths such that each LSR along the path that must load-balance
capabilities in accessing the depth of a label stack. The label is able to access at least one EL. Note that the source LSR
stack depth access value of intermediate LSRs must be known to create may have to insert multiple ELs in the label stack at different depths
such a label stack. How this value is determined is outside the for this to work since intermediate LSRs may have differing
scope of this document. This value can be advertised using a capabilities in accessing the depth of a label stack. The label stack
protocol such as an IGP. depth access value of intermediate LSRs must be known to create such a
</t> label stack. How this value is determined is outside the scope of
<t> this document. This value can be advertised using a protocol such as
Applying this method to the example in <xref target="usecase"/> above, an IGP.
if LSR P1 </t>
needs to have the EL within a depth of 4, then the source LSR S <t pn="section-10.5-2">
encoded label stack would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI, Applying this method to the example in <xref target="usecase" format="
EL&gt; where all the ELs would typically have the same value. default" sectionFormat="of" derivedContent="Section 3"/>, if LSR P1
</t> needs to have the EL within a depth of 4, then the source label stack that
<t> is LSR S encoded would be &lt;L_N-P3, ELI, EL, L_A-L1, L_N-D, ELI,
EL&gt;, where all the ELs would typically have the same value.
</t>
<t pn="section-10.5-3">
In the case where the ERLD has different values along the path and the In the case where the ERLD has different values along the path and the
LSR that is inserting &lt;ELI, EL&gt; pairs has no limit on how many pairs LSR that is inserting &lt;ELI, EL&gt; pairs has no limit on how many pairs
it can insert, and it knows the appropriate positions in the stack it can insert, and it knows the appropriate positions in the stack
where they should be inserted, this option is the same as the where they should be inserted, this option is the same as the
recommended solution in <xref target="solution"/>. recommended solution in <xref target="solution" format="default" sectionF
</t> ormat="of" derivedContent="Section 7"/>.
<t> </t>
Note that a refinement of this solution which balances the number of <t pn="section-10.5-4">
pushed labels against the desired entropy is the solution described Note that a refinement of this solution, which balances the number of
in <xref target="solution"/>. pushed labels against the desired entropy, is the solution described
</t> in <xref target="solution" format="default" sectionFormat="of" derivedContent
</section> ="Section 7"/>.
</section> </t>
</section>
<section title="Acknowledgements"> </section>
<t>The authors would like to thank John Drake, Loa Andersson, Curtis <section toc="include" numbered="true" removeInRFC="false" pn="section-11">
Villamizar, Greg Mirsky, Markus Jork, Kamran Raza, Carlos Pignataro, Bruno De <name slugifiedName="name-iana-considerations">IANA Considerations</name>
craene, Chris Bowers, Nobo Akiya, Daniele Ceccarelli and Joe Clarke for <t pn="section-11-1"> This document has no IANA actions.
their review comments and suggestions.
</t>
</section>
<section title="Contributors">
<figure>
<artwork>
Xiaohu Xu
Huawei
Email: xuxiaohu@huawei.com
Wim Hendrickx
Nokia
Email: wim.henderickx@nokia.com
Gunter Van De Velde
Nokia
Email: gunter.van_de_velde@nokia.com
Acee Lindem
Cisco
Email: acee@cisco.com
</artwork>
</figure>
</section>
<section title="IANA Considerations" toc="default">
<t> This memo includes no request to IANA. Note to RFC Editor: Remove
this section before publication.
</t> </t>
</section> </section>
<section title="Security Considerations" toc="default"> <section toc="include" numbered="true" removeInRFC="false" pn="section-12">
<t>Compared to <xref target="RFC6790"/>, this document introduces the notion of <name slugifiedName="name-security-considerations">Security Considerations
ERLD, MSD and may require an ingress node to push multiple ELI/EL. </name>
These changes does not introduce any new security considerations <t pn="section-12-1">Compared to <xref target="RFC6790" format="default" s
beyond those already listed in <xref target="RFC6790"/>. ectionFormat="of" derivedContent="RFC6790"/>, this document introduces the notio
n
of ERLD and MSD, and may require an ingress node to push multiple ELIs/ELs.
These changes do not introduce any new security considerations beyond those
already listed in <xref target="RFC6790" format="default" sectionFormat="of" der
ivedContent="RFC6790"/>.
</t> </t>
</section> </section>
</middle> </middle>
<back> <back>
<references title="Normative References"> <displayreference target="I-D.ietf-idr-bgp-ls-segment-routing-msd" to="MSD-B
&RFC2119; GP"/>
&RFC6790; <displayreference target="I-D.ietf-isis-mpls-elc" to="ISIS-ELC"/>
&RFC8174; <displayreference target="I-D.ietf-ospf-mpls-elc" to="OSPF-ELC"/>
&SR; <references pn="section-13">
&SR-MPLS; <name slugifiedName="name-references">References</name>
</references> <references pn="section-13.1">
<references title="Informative References"> <name slugifiedName="name-normative-references">Normative References</na
&ISIS-ELC; me>
&OSPF-ELC; <reference anchor="RFC6790" target="https://www.rfc-editor.org/info/rfc6
&SR-L2-BUNDLES; 790" quoteTitle="true" derivedAnchor="RFC6790">
&RFC7855; <front>
&ISIS-MSD; <title>The Use of Entropy Labels in MPLS Forwarding</title>
&OSPF-MSD; <author initials="K." surname="Kompella" fullname="K. Kompella">
&BGP-MSD; <organization showOnFrontPage="true"/>
</author>
<author initials="J." surname="Drake" fullname="J. Drake">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Amante" fullname="S. Amante">
<organization showOnFrontPage="true"/>
</author>
<author initials="W." surname="Henderickx" fullname="W. Henderickx">
<organization showOnFrontPage="true"/>
</author>
<author initials="L." surname="Yong" fullname="L. Yong">
<organization showOnFrontPage="true"/>
</author>
<date year="2012" month="November"/>
<abstract>
<t>Load balancing is a powerful tool for engineering traffic acros
s a network. This memo suggests ways of improving load balancing across MPLS ne
tworks using the concept of "entropy labels". It defines the concept, describes
why entropy labels are useful, enumerates properties of entropy labels that all
ow maximal benefit, and shows how they can be signaled and used for various appl
ications. This document updates RFCs 3031, 3107, 3209, and 5036. [STANDARDS-TR
ACK]</t>
</abstract>
</front>
<seriesInfo name="RFC" value="6790"/>
<seriesInfo name="DOI" value="10.17487/RFC6790"/>
</reference>
<reference anchor="RFC2119" target="https://www.rfc-editor.org/info/rfc2
119" quoteTitle="true" derivedAnchor="RFC2119">
<front>
<title>Key words for use in RFCs to Indicate Requirement Levels</tit
le>
<author initials="S." surname="Bradner" fullname="S. Bradner">
<organization showOnFrontPage="true"/>
</author>
<date year="1997" month="March"/>
<abstract>
<t>In many standards track documents several words are used to sig
nify the requirements in the specification. These words are often capitalized.
This document defines these words as they should be interpreted in IETF document
s. This document specifies an Internet Best Current Practices for the Internet
Community, and requests discussion and suggestions for improvements.</t>
</abstract>
</front>
<seriesInfo name="BCP" value="14"/>
<seriesInfo name="RFC" value="2119"/>
<seriesInfo name="DOI" value="10.17487/RFC2119"/>
</reference>
<reference anchor="RFC8174" target="https://www.rfc-editor.org/info/rfc8
174" quoteTitle="true" derivedAnchor="RFC8174">
<front>
<title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</ti
tle>
<author initials="B." surname="Leiba" fullname="B. Leiba">
<organization showOnFrontPage="true"/>
</author>
<date year="2017" month="May"/>
<abstract>
<t>RFC 2119 specifies common key words that may be used in protoco
l specifications. This document aims to reduce the ambiguity by clarifying tha
t only UPPERCASE usage of the key words have the defined special meanings.</t>
</abstract>
</front>
<seriesInfo name="BCP" value="14"/>
<seriesInfo name="RFC" value="8174"/>
<seriesInfo name="DOI" value="10.17487/RFC8174"/>
</reference>
<reference anchor="RFC8402" target="https://www.rfc-editor.org/info/rfc8
402" quoteTitle="true" derivedAnchor="RFC8402">
<front>
<title>Segment Routing Architecture</title>
<author initials="C." surname="Filsfils" fullname="C. Filsfils" role
="editor">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Previdi" fullname="S. Previdi" role="
editor">
<organization showOnFrontPage="true"/>
</author>
<author initials="L." surname="Ginsberg" fullname="L. Ginsberg">
<organization showOnFrontPage="true"/>
</author>
<author initials="B." surname="Decraene" fullname="B. Decraene">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Litkowski" fullname="S. Litkowski">
<organization showOnFrontPage="true"/>
</author>
<author initials="R." surname="Shakir" fullname="R. Shakir">
<organization showOnFrontPage="true"/>
</author>
<date year="2018" month="July"/>
<abstract>
<t>Segment Routing (SR) leverages the source routing paradigm. A
node steers a packet through an ordered list of instructions, called "segments".
A segment can represent any instruction, topological or service based. A segm
ent can have a semantic local to an SR node or global within an SR domain. SR p
rovides a mechanism that allows a flow to be restricted to a specific topologica
l path, while maintaining per-flow state only at the ingress node(s) to the SR d
omain.</t>
<t>SR can be directly applied to the MPLS architecture with no cha
nge to the forwarding plane. A segment is encoded as an MPLS label. An ordered
list of segments is encoded as a stack of labels. The segment to process is on
the top of the stack. Upon completion of a segment, the related label is poppe
d from the stack.</t>
<t>SR can be applied to the IPv6 architecture, with a new type of
routing header. A segment is encoded as an IPv6 address. An ordered list of se
gments is encoded as an ordered list of IPv6 addresses in the routing header. T
he active segment is indicated by the Destination Address (DA) of the packet. T
he next active segment is indicated by a pointer in the new routing header.</t>
</abstract>
</front>
<seriesInfo name="RFC" value="8402"/>
<seriesInfo name="DOI" value="10.17487/RFC8402"/>
</reference>
<reference anchor="RFC8660" target="https://www.rfc-editor.org/info/rfc8
660" quoteTitle="true" derivedAnchor="RFC8660">
<front>
<title>Segment Routing with the MPLS Data Plane</title>
<author initials="A" surname="Bashandy" fullname="Ahmed Bashandy" ro
le="editor">
<organization showOnFrontPage="true"/>
</author>
<author initials="C" surname="Filsfils" fullname="Clarence" role="ed
itor">
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Previdi" fullname="Stefano Previdi">
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Litkowski" fullname="Stephane Litkowsk
i">
<organization showOnFrontPage="true"/>
</author>
<author initials="R" surname="Shakir" fullname="Rob Shakir">
<organization showOnFrontPage="true"/>
</author>
<date month="December" year="2019"/>
</front>
<seriesInfo name="RFC" value="8660"/>
<seriesInfo name="DOI" value="10.17487/RFC8660"/>
</reference>
</references>
<references pn="section-13.2">
<name slugifiedName="name-informative-references">Informative References
</name>
<reference anchor="I-D.ietf-isis-mpls-elc" quoteTitle="true" target="htt
ps://tools.ietf.org/html/draft-ietf-isis-mpls-elc-10" derivedAnchor="ISIS-ELC">
<front>
<title>Signaling Entropy Label Capability and Entropy Readable Label
Depth Using IS-IS</title>
<author initials="X" surname="Xu" fullname="Xiaohu Xu">
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Kini" fullname="Sriganesh Kini">
<organization showOnFrontPage="true"/>
</author>
<author initials="P" surname="Psenak" fullname="Peter Psenak">
<organization showOnFrontPage="true"/>
</author>
<author initials="C" surname="Filsfils" fullname="Clarence Filsfils"
>
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Litkowski" fullname="Stephane Litkowsk
i">
<organization showOnFrontPage="true"/>
</author>
<author initials="M" surname="Bocci" fullname="Matthew Bocci">
<organization showOnFrontPage="true"/>
</author>
<date month="October" day="21" year="2019"/>
<abstract>
<t>Multiprotocol Label Switching (MPLS) has defined a mechanism to
load- balance traffic flows using Entropy Labels (EL). An ingress Label Switch
ing Router (LSR) cannot insert ELs for packets going into a given Label Switched
Path (LSP) unless an egress LSR has indicated via signaling that it has the cap
ability to process ELs, referred to as Entropy Label Capability (ELC), on that t
unnel. In addition, it would be useful for ingress LSRs to know each LSR's capa
bility for reading the maximum label stack depth and performing EL-based load- b
alancing, referred to as Entropy Readable Label Depth (ERLD). This document def
ines a mechanism to signal these two capabilities using IS-IS. These mechanisms
are particularly useful, where label advertisements are done via protocols like
IS-IS.</t>
</abstract>
</front>
<seriesInfo name="Internet-Draft" value="draft-ietf-isis-mpls-elc-10"/
>
<format type="TXT" target="http://www.ietf.org/internet-drafts/draft-i
etf-isis-mpls-elc-10.txt"/>
<refcontent>Work in Progress</refcontent>
</reference>
<reference anchor="I-D.ietf-ospf-mpls-elc" quoteTitle="true" target="htt
ps://tools.ietf.org/html/draft-ietf-ospf-mpls-elc-12" derivedAnchor="OSPF-ELC">
<front>
<title>Signaling Entropy Label Capability and Entropy Readable Label
-stack Depth Using OSPF</title>
<author initials="X" surname="Xu" fullname="Xiaohu Xu">
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Kini" fullname="Sriganesh Kini">
<organization showOnFrontPage="true"/>
</author>
<author initials="P" surname="Psenak" fullname="Peter Psenak">
<organization showOnFrontPage="true"/>
</author>
<author initials="C" surname="Filsfils" fullname="Clarence Filsfils"
>
<organization showOnFrontPage="true"/>
</author>
<author initials="S" surname="Litkowski" fullname="Stephane Litkowsk
i">
<organization showOnFrontPage="true"/>
</author>
<author initials="M" surname="Bocci" fullname="Matthew Bocci">
<organization showOnFrontPage="true"/>
</author>
<date month="October" day="25" year="2019"/>
<abstract>
<t>Multiprotocol Label Switching (MPLS) has defined a mechanism to
load- balance traffic flows using Entropy Labels (EL). An ingress Label Switch
ing Router (LSR) cannot insert ELs for packets going into a given tunnel unless
an egress LSR has indicated via signaling that it has the capability to process
ELs, referred to as Entropy Label Capability (ELC), on that tunnel. In addition
, it would be useful for ingress LSRs to know each LSR's capability of reading t
he maximum label stack depth and performing EL-based load-balancing, referred to
as Entropy Readable Label Depth (ERLD). This document defines a mechanism to s
ignal these two capabilities using OSPF and OSPFv3. These mechanism is particula
rly useful in the environment where Segment Routing (SR) is used, where label ad
vertisements are done via protocols like OSPF and OSPFv3.</t>
</abstract>
</front>
<seriesInfo name="Internet-Draft" value="draft-ietf-ospf-mpls-elc-12"/
>
<format type="TXT" target="http://www.ietf.org/internet-drafts/draft-i
etf-ospf-mpls-elc-12.txt"/>
<refcontent>Work in Progress</refcontent>
</reference>
<reference anchor="RFC8668" target="https://www.rfc-editor.org/info/rfc8
668" quoteTitle="true" derivedAnchor="RFC8668">
<front>
<title>Advertising Layer 2 Bundle Member Link Attributes in IS-IS</t
itle>
<author initials="L" surname="Ginsberg" fullname="Les Ginsberg">
<organization showOnFrontPage="true"/>
</author>
<author initials="A" surname="Bashandy" fullname="Ahmed Bashandy">
<organization showOnFrontPage="true"/>
</author>
<author initials="C" surname="Filsfils" fullname="Clarence Filsfils"
>
<organization showOnFrontPage="true"/>
</author>
<author initials="M" surname="Nanduri" fullname="Mohan Nanduri">
<organization showOnFrontPage="true"/>
</author>
<author initials="E" surname="Aries" fullname="Ebben Aries">
<organization showOnFrontPage="true"/>
</author>
<date month="December" year="2019"/>
</front>
<seriesInfo name="RFC" value="8668"/>
<seriesInfo name="DOI" value="10.17487/RFC8668"/>
</reference>
<reference anchor="RFC7855" target="https://www.rfc-editor.org/info/rfc7
855" quoteTitle="true" derivedAnchor="RFC7855">
<front>
<title>Source Packet Routing in Networking (SPRING) Problem Statemen
t and Requirements</title>
<author initials="S." surname="Previdi" fullname="S. Previdi" role="
editor">
<organization showOnFrontPage="true"/>
</author>
<author initials="C." surname="Filsfils" fullname="C. Filsfils" role
="editor">
<organization showOnFrontPage="true"/>
</author>
<author initials="B." surname="Decraene" fullname="B. Decraene">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Litkowski" fullname="S. Litkowski">
<organization showOnFrontPage="true"/>
</author>
<author initials="M." surname="Horneffer" fullname="M. Horneffer">
<organization showOnFrontPage="true"/>
</author>
<author initials="R." surname="Shakir" fullname="R. Shakir">
<organization showOnFrontPage="true"/>
</author>
<date year="2016" month="May"/>
<abstract>
<t>The ability for a node to specify a forwarding path, other than
the normal shortest path, that a particular packet will traverse, benefits a nu
mber of network functions. Source-based routing mechanisms have previously been
specified for network protocols but have not seen widespread adoption. In this
context, the term "source" means "the point at which the explicit route is impo
sed"; therefore, it is not limited to the originator of the packet (i.e., the no
de imposing the explicit route may be the ingress node of an operator's network)
.</t>
<t>This document outlines various use cases, with their requiremen
ts, that need to be taken into account by the Source Packet Routing in Networkin
g (SPRING) architecture for unicast traffic. Multicast use cases and requiremen
ts are out of scope for this document.</t>
</abstract>
</front>
<seriesInfo name="RFC" value="7855"/>
<seriesInfo name="DOI" value="10.17487/RFC7855"/>
</reference>
<reference anchor="RFC8476" target="https://www.rfc-editor.org/info/rfc8
476" quoteTitle="true" derivedAnchor="RFC8476">
<front>
<title>Signaling Maximum SID Depth (MSD) Using OSPF</title>
<author initials="J." surname="Tantsura" fullname="J. Tantsura">
<organization showOnFrontPage="true"/>
</author>
<author initials="U." surname="Chunduri" fullname="U. Chunduri">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Aldrin" fullname="S. Aldrin">
<organization showOnFrontPage="true"/>
</author>
<author initials="P." surname="Psenak" fullname="P. Psenak">
<organization showOnFrontPage="true"/>
</author>
<date year="2018" month="December"/>
<abstract>
<t>This document defines a way for an Open Shortest Path First (OS
PF) router to advertise multiple types of supported Maximum SID Depths (MSDs) at
node and/or link granularity. Such advertisements allow entities (e.g., centra
lized controllers) to determine whether a particular Segment Identifier (SID) st
ack can be supported in a given network. This document only refers to the Signa
ling MSD as defined in RFC 8491, but it defines an encoding that can support oth
er MSD types. Here, the term "OSPF" means both OSPFv2 and OSPFv3.</t>
</abstract>
</front>
<seriesInfo name="RFC" value="8476"/>
<seriesInfo name="DOI" value="10.17487/RFC8476"/>
</reference>
<reference anchor="RFC8491" target="https://www.rfc-editor.org/info/rfc8
491" quoteTitle="true" derivedAnchor="RFC8491">
<front>
<title>Signaling Maximum SID Depth (MSD) Using IS-IS</title>
<author initials="J." surname="Tantsura" fullname="J. Tantsura">
<organization showOnFrontPage="true"/>
</author>
<author initials="U." surname="Chunduri" fullname="U. Chunduri">
<organization showOnFrontPage="true"/>
</author>
<author initials="S." surname="Aldrin" fullname="S. Aldrin">
<organization showOnFrontPage="true"/>
</author>
<author initials="L." surname="Ginsberg" fullname="L. Ginsberg">
<organization showOnFrontPage="true"/>
</author>
<date year="2018" month="November"/>
<abstract>
<t>This document defines a way for an Intermediate System to Inter
mediate System (IS-IS) router to advertise multiple types of supported Maximum S
ID Depths (MSDs) at node and/or link granularity. Such advertisements allow enti
ties (e.g., centralized controllers) to determine whether a particular Segment I
D (SID) stack can be supported in a given network. This document only defines o
ne type of MSD: Base MPLS Imposition. However, it defines an encoding that can
support other MSD types. This document focuses on MSD use in a network that is
Segment Routing (SR) enabled, but MSD may also be useful when SR is not enabled.
</t>
</abstract>
</front>
<seriesInfo name="RFC" value="8491"/>
<seriesInfo name="DOI" value="10.17487/RFC8491"/>
</reference>
<reference anchor="I-D.ietf-idr-bgp-ls-segment-routing-msd" quoteTitle="
true" target="https://tools.ietf.org/html/draft-ietf-idr-bgp-ls-segment-routing-
msd-09" derivedAnchor="MSD-BGP">
<front>
<title>Signaling MSD (Maximum SID Depth) using Border Gateway Protoc
ol Link-State</title>
<author initials="J" surname="Tantsura" fullname="Jeff Tantsura">
<organization showOnFrontPage="true"/>
</author>
<author initials="U" surname="Chunduri" fullname="Uma Chunduri">
<organization showOnFrontPage="true"/>
</author>
<author initials="K" surname="Talaulikar" fullname="Ketan Talaulikar
">
<organization showOnFrontPage="true"/>
</author>
<author initials="G" surname="Mirsky" fullname="Gregory Mirsky">
<organization showOnFrontPage="true"/>
</author>
<author initials="N" surname="Triantafillis" fullname="Nikos Trianta
fillis">
<organization showOnFrontPage="true"/>
</author>
<date month="October" day="15" year="2019"/>
<abstract>
<t>This document defines a way for a Border Gateway Protocol Link-
State (BGP-LS) speaker to advertise multiple types of supported Maximum SID Dept
hs (MSDs) at node and/or link granularity. Such advertisements allow entities (
e.g., centralized controllers) to determine whether a particular Segment Identif
ier (SID) stack can be supported in a given network.</t>
</abstract>
</front>
<seriesInfo name="Internet-Draft" value="draft-ietf-idr-bgp-ls-segment
-routing-msd-09"/>
<format type="TXT" target="http://www.ietf.org/internet-drafts/draft-i
etf-idr-bgp-ls-segment-routing-msd-09.txt"/>
<refcontent>Work in Progress</refcontent>
</reference>
</references>
</references> </references>
<section numbered="false" toc="include" removeInRFC="false" pn="section-appe
ndix.a">
<name slugifiedName="name-acknowledgements">Acknowledgements</name>
<t pn="section-appendix.a-1">The authors would like to thank John Drake, L
oa Andersson, Curtis
Villamizar, Greg Mirsky, Markus Jork, Kamran Raza, Carlos Pignataro, Bruno De
craene, Chris Bowers, Nobo Akiya, Daniele Ceccarelli, and Joe Clarke for
their review, comments, and suggestions.
</t>
</section>
<section numbered="false" toc="include" removeInRFC="false" pn="section-appe
ndix.b">
<name slugifiedName="name-contributors">Contributors</name>
<artwork name="" type="" align="left" alt="" pn="section-appendix.b-1">
Xiaohu Xu
Huawei
Email: xuxiaohu@huawei.com
</artwork>
<artwork name="" type="" align="left" alt="" pn="section-appendix.b-2">
Wim Hendrickx
Nokia
Email: wim.henderickx@nokia.com
</artwork>
<artwork name="" type="" align="left" alt="" pn="section-appendix.b-3">
Gunter Van de Velde
Nokia
Email: gunter.van_de_velde@nokia.com
</artwork>
<artwork name="" type="" align="left" alt="" pn="section-appendix.b-4">
Acee Lindem
Cisco
Email: acee@cisco.com
</artwork>
</section>
<section anchor="authors-addresses" numbered="false" removeInRFC="false" toc
="include" pn="section-appendix.c">
<name slugifiedName="name-authors-addresses">Authors' Addresses</name>
<author initials="S" surname="Kini" fullname="Sriganesh Kini">
<organization showOnFrontPage="true"/>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>sriganeshkini@gmail.com</email>
</address>
</author>
<author initials="K" surname="Kompella" fullname="Kireeti Kompella">
<organization showOnFrontPage="true">Juniper</organization>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>kireeti@juniper.net</email>
</address>
</author>
<author initials="S" surname="Sivabalan" fullname="Siva Sivabalan">
<organization showOnFrontPage="true">Cisco</organization>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>msiva@cisco.com</email>
</address>
</author>
<author initials="S" surname="Litkowski" fullname="Stephane Litkowski">
<organization showOnFrontPage="true">Orange</organization>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>slitkows.ietf@gmail.com</email>
</address>
</author>
<author initials="R" surname="Shakir" fullname="Rob Shakir">
<organization showOnFrontPage="true">Google</organization>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>robjs@google.com</email>
</address>
</author>
<author initials="J" surname="Tantsura" fullname="Jeff Tantsura">
<organization showOnFrontPage="true">Apstra, Inc.</organization>
<address>
<postal>
<street/>
<city/>
<region/>
<code/>
<country/>
</postal>
<email>jefftant.ietf@gmail.com</email>
</address>
</author>
</section>
</back> </back>
</rfc> </rfc>
 End of changes. 89 change blocks. 
1046 lines changed or deleted 1930 lines changed or added

This html diff was produced by rfcdiff 1.45. The latest version is available from http://tools.ietf.org/tools/rfcdiff/