Segment Routing-Technology & Use Cases

Segment Routing: Technology
Update and Advanced Use-Cases

Steve Braaten, Solutions Architect
BRKRST-3122
Agenda
Segment Routing Executive Summary
Reminders
Incremental Deployment Use-Cases
Inter-Domain Policy at Scale

Topology Independent LFA (TI-LFA)
Microloop Avoidance
Conclusion
Segment Routing
Executive Summary
Segment Routing
Source Routing
the source chooses a path and encodes it in the packet header as an ordered list of
segments
the rest of the network executes the encoded instructions
Segment: an identifier for any type of instruction
forwarding or service
Forwarding Plane:
MPLS: an ordered list of segments is represented as a stack of labels

IPv6: an ordered list of segments is encoded in a routing extension header
Multi-Vendor solution
BRKRST-3122
2016 Cisco and/or its affiliates. All rights reserved. Cisco Public
Deployed !
First deployments in 2015 just 15 months after FCS !!!
Strong start in 2016 with many new deployments
BRKRST-3122
www.segment-routing.net
tools.ietf.org/wg/spring/
IETF
Strong commitment for standardization and

multi-vendor support
SPRING Working-Group (started Nov 2013)
All key documents are WG-status
Over 25 drafts maintained by SR team
Over 50% are WG status

Over 75% have a Cisco implementation
Several interop reports are available
First RFC document - RFC 7855 (May 2016)
BRKRST-3122
Foundation for modern IP/MPLS networking
Simplicity
Solution to unsolved problems
End-to-end policy, local and/or centralized PCE, 50msec protection, microloop

avoidance, and more
Scale
Set of few, well-chosen building blocks
Granular traffic engineering with minimal network state
Seamless Deployment
SR/LDP interworking, SR/RSVP-TE interworking, ship-in-the-night co-existence
Decoupled data and control planes
Low-cost
BRKRST-3122
Session Assumptions
Thorough understanding of SR tutorial
You should know
http://www.segment-routing.net/home/tutorial
SRGB
IGP Prefix and Adj SIDs
Anycast SID
SR/LDP interworking
BGP Prefix SID and the MSDC use-case
Use-Cases in this presentation described for SR/MPLS
Same concept applies to SRv6 (native IPv6 SR extension header, no MPLS)
BRKRST-3122
Reminders
Prefix segment
16004
Shortest-path to the prefix

Equal Cost MultiPath (ECMP)-aware
Global Segment
2
16004
Label = 16000 + Index
16004
16004
Advertised as index
16004
16004
Distributed by ISIS/OSPF/BGP
4
1.1.1.4/32
16004
All nodes use default SRGB
16,000 23,999
BRKRST-3122
11
Adjacency segment
Forward on the IGP adjacency
Local Segment
Advertised as label value
Adj to 5
24025
2
Adj to 4
24024
Distributed by ISIS/OSPF
But only local adjacency SIDs are

installed in FIB!
All nodes use default SRGB

16,000 23,999
BRKRST-3122
12
SR operations illustration
Goal: Go to Z
but avoid node 7
16005
16005
56056
56056
16101
16101
Payload to Z
Payload to Z
Steer traffic on any path through the

network
Path is specified by list of segments

in packet header, a stack of labels
No path is signaled
No per-flow state is created
IS-IS, OSPF, BGP all supported
56056
16101
16101
Payload to Z
56056
101
Node Z
Payload to Z
Payload to Z
16101
Payload to Z
BRKRST-3122
13
Incremental Deployment
Use-Cases
SR Innovation
VPN /
Service
Transport
Data Center
Fabric
Egress
Peering
Engineering
Topology
Independent
IP FRR
Incremental
Use Case
Deployment
Microloop
Avoidance
Demand
Matrix
Inter-Domain
Policy at
Scale
Application
Engineered
Routing
Traffic
Engineering
http://blogs.cisco.com/sp/supercharge-your-network-with-segment-routing-innovations
BRKRST-3122
15
SR Innovation
VPN /
Service
Transport
Data Center
Fabric
Egress
Peering
Engineering
Topology
Independent
IP FRR
Incremental
Use Case
Deployment
Microloop
Avoidance
Demand
Matrix
Inter-Domain
Policy at
Scale
Application
Engineered
Routing
Traffic
Engineering
http://blogs.cisco.com/sp/supercharge-your-network-with-segment-routing-innovations
BRKRST-3122
16

draft-filsfils-spring-large-scale-interconnect
Use-Case Description
DC A1
vPE1
ToR
Spine
METRO A
LSR
WAN
LSR
METRO B
Spine
LSR
ToR
vPE2
Datacenter
Core
Datacenter
DC B2
Metro
Metro
Segment Routing use-case aiming to scale the network to support hundreds of

thousands of network nodes, and tens of millions of physical underlay endpoints
Applicable to the interconnection of massive-scale DC's and/or large

aggregation networks
Principles are equally applicable to a network of any size

BRKRST-3122
18
SRGB and SID allocation

DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
DC B2
Spine
20003
LSR
18002
17k-18k
vPE2
20001
20k-24k
16k-17k
20k-24k
ToR
20002
18k-19k
16k-24k
Homogenous end-to-end SRGB for simplicity
Globally Unique Prefix SIDs for devices WAN and Metro domains
Locally Unique Prefix SIDs for Datacenters

BRKRST-3122
19
IGP/SR within WAN and Metro Domains

METRO A
DCI1
17001
LSR
17002
IGP / SR 2
WAN
LSR
16003
IGP / SR 1
METRO B
LSR
18002
DCI2
18001
IGP / SR 3
Each domain runs ISIS/OSPF SR
Incremental deployment and seamless interworking with LDP
BRKRST-3122
20
Segment Routing in the Datacenter

AS1
vPE1
20001
AS2
AS3
AS4
AS5
ToR2
20002
Leaf3
20003
Spine4
ToR12
20012
Leaf13
20013
Spine14
20004
AS6
Leaf5
20005
DCI6
20006
Leaf15
20015
DCI16
20016
AS11
vPE11
20011
20014
Datacenter fabric runs BGP SR
Example: 20006 is the BGP Prefix SID to DCI6
ECMP-aware
Simple (no LDP/RSVP)
Policy-driven
BRKRST-3122
21

SR connectivity across domains
Intra-Domain Routing DC A1 and DC B2

DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
LSR
18002
vPE1/32
NH: vPE1
BGP-LU LABEL: POP
PREFIX-SID: 20001
(relative 4001)
Spine
20003
ToR
20002
vPE2
20001
DCI2/32
NH: DCI2
BGP-LU LABEL: POP
PREFIX-SID: 18001
(relative 2001)
vPE1/32
NH: TOR
BGP-LU LABEL: 20001
PREFIX-SID: 20001
(relative 4001)
DCI2/32
NH: SPINE
BGP-LU LABEL: 18001
PREFIX-SID: 18001
(relative 2001)
vPE1/32
NH: SPINE
BGP-LU LABEL: 20001
PREFIX-SID: 20001
(relative 4001)
DC B2
DCI2/32
NH: TOR
BGP-LU LABEL: 18001
PREFIX-SID: 18001
(relative 2001)
BGP SR in the DC
Often eBGP would be used but iBGP can also be used (see tutorial)
Smart AS (ClusterID) allocation in eBGP (iBGP) provides automated path filtering (see
tutorial)
BRKRST-3122
23
Intra-Domain Routing Metro A and Metro B

DC A1
vPE1
20001
ToR
20002
METRO A
Spine
20003
LSR
17002
DCI1/32
NH: DCI1
BGP-LU LABEL: POP
PREFIX-SID: 17001
(relative 1001)
WAN
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
ISIS LSP of AGG2

Leaf: Agg2
PREFIX-SID: 16002
(relative 2)
DCI1/32
NH: LSR
BGP-LU LABEL: 17001
PREFIX-SID: 17001
(relative 1001)
In a metro, BGP/SR or ISIS-OSPF/SR are likely, both illustrated here
Example: Metro A: BGP/SR

Example: Metro B: ISIS/SR
BRKRST-3122
24
Intra-Domain Routing WAN

DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
WAN
LSR
17002
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
ISIS LSP of AGG1

Leaf: Agg1
PREFIX-SID: 16001
(relative 1)
ISIS / OSPF SR in WAN
During a migration, benefit from SR seamless interworking with LDP and ship-inthe-night with RSVP
BRKRST-3122
25
Inter-Domain Routing
DC A1
vPE1
20001
ToR
20002
Spine
20003
WAN Aggs
METRO A
LSR
17002
WAN Aggs
WAN
LSR
16003
METRO B
Spine
20003
LSR
18002
WAN Aggs
WAN aggs are re-distributed down to Metro and DC
Nothing is redistributed up !!!
How does vPE1 reaches vPE2?
BRKRST-3122
DC B2
ToR
20002
vPE2
20001
WAN Aggs
26
Inter-Domain Routing
DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
DC B2
Spine
20003
LSR
18002
ToR
20002
vPE2
20001
ISIS LSP of AGG2

Leaf: AGG1
PREFIX-SID: 16001
(relative 1)
AGG1
NH: DCI2
BGP-LU LABEL: 16001
PREFIX-SID: 16001
(relative 1)
Redistribution: from center to leaves

WAN redistributes (only) its AGGs into metros
Metro redistributes (only) the WAN AGGs into DCs
Redistribution: from leaves to center
AGG1
NH: SPINE
BGP-LU LABEL: 16001
PREFIX-SID: 16001
(relative 1)
AGG1
NH: ToR
BGP-LU LABEL: 16001
PREFIX-SID: 16001
(relative 1)
Nothing
BRKRST-3122
27
Inter-Domain Routing (Contd)

DC A1
vPE1
20001
ToR
20002
METRO A
Spine
20003
WAN
LSR
17002
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
ISIS LSP of AGG1

Leaf: AGG2
PREFIX-SID: 16002
(relative 2)
AGG2
NH: DCI1
BGP-LU LABEL: 16002
PREFIX-SID: 16002
(relative 2)
AGG2
NH: SPINE
BGP-LU LABEL: 16002
PREFIX-SID: 16002
(relative 2)
AGG2
NH: TOR
BGP-LU LABEL: 16002
PREFIX-SID: 16002
(relative 2)
BRKRST-3122
28

SR PCE
SR PCE
Demo
Compute
SR PCE
Multi-Domain Topology
DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
Multi-Domain topology
Real-time reactive feed via BGP-LS/ISIS/OSPF from multiple domains

Including ip address and SID
Compute: stateful with native SRTE algorithms
BRKRST-3122
30
Circuit Optimization vs SR Optimization

2
5
4
1
4
7
6
7
6
Classic TE is circuit-based
CSPF => non-ECMP path
SID List: {4, 5, 7, 3}
Poor ECMP, big SR list, ATM optimized
SR-native TE algorithms needed

Recognized Innovation - Sigcomm 2015
SID List: {7, 3}
ECMP, Small SR list, IP-optimized
BRKRST-3122
31
Four SR-native TE algorithms developed
Metric optimization with inclusion/exclusion constraint and bound
Metric: IGP metric, TE metric, extended TE-latency metric

Inclusion/exclusion: IP address, SRLG, TE affinity, Link Loss
Margin: any solution within the margin of the optimum is accepted
Favor more ECMP or shorter SID list instead of insignificant optimization increment
Also available on the router-based SRTE functionality
Disjointness
(A to Z) or ((A, B) to (Y, Z))

With minimized latency diff, ECMP and shorter SID list
(A to Z) also available on the router-based SRTE functionality
Tactical BW optimization
Multi-Constrained
Sigcomm 2015 [url]
Furthermore TI-LFA and Microloop avoidance algorithms

BRKRST-3122
32
SR PCE - Fundamentally Distributed

DC A1
vPE1
20001
ToR
20002
Spine
20003
WAN
METRO A
SR
PCE
LSR
17002
SR
PCE
SR
PCE
SR
PCE
LSR
16003
METRO B
SR
PCE
LSR
18002
DC B2
SR
PCE
Spine
20003
ToR
20002
vPE2
20001
SR
PCE
SR
PCE
SR PCE not to be considered as a single God box
SR PCE deployment model more like BGP Route Reflectors
Different vPEs can use different pairs of SR PCEs
SR PCE preference can either be based on proximity or service
BRKRST-3122
33

On-Demand SR Next Hop (ODN)
Service Provisioning
DC A1
vPE1
20001
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
vPE1 learns about a service route with nhop vPE2
ToR
20002
1: V via vPE2
VPN-LABEL: 99999
BGP
RR
2: V via vPE2
VPN-LABEL: 99999
RR shown could be any flavor of overlay controller
How does vPE1 reach the nhop?

vPE1 only has routes within DC A1 and to the AGGs of the WAN domain
Solution: On-Demand SR Next Hop (ODN)
BRKRST-3122
35
On-Demand SR Next-Hop
Demo
Overview
4: {SID List}
2: V via vPE2
VPN-LABEL: 99999
1: V via vPE2
VPN-LABEL: 99999
BGP
RR
SR
PCE
DC A1
3: vPE2 ?
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
When the vPEs does not have any RIB entry for the (locator, policy), the On-Demand SR Next-Hop
automatically sends a stateful PCEP request to the SR PCE
Key benefit: provide the glue between the overlay and underlay controllers while decoupling them
E.g. overlay controller does not need to react to multi-domain underlay topology change, nor compute TE policies
E.g. underlay controller does not need to be involved in service orchestration, does not store any a priori TE policy
E.g. no direct API or coupled workflow between the controllers
BRKRST-3122
36
Demo
Reachability
4: {16002, 18001, 20001}
2: V via vPE2
VPN-LABEL: 99999
1: V via vPE2
VPN-LABEL: 99999
BGP
RR
SR
PCE
DC A1
3: vPE2 ?
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
LSR
16003
METRO B
16002
LSR
18002
18001
DC B2
Spine
20003
ToR
20002
vPE1s ODN functionality automatically request a solution from SR PCE
Scalable: vPE1 only gets the inter-domain paths that it needs
Simple: no BGP3107 pushing all routes everywhere
BRKRST-3122
vPE2
20001
37
Demo
SLA enabled
4: {16001, 16003,
16002, 18001, 20001}
2: V via vPE2
VPN-LABEL: 99999
EXT-COM: LATENCY
1: V via vPE2
VPN-LABEL: 99999
EXT-COM: LATENCY
BGP
RR
SR
PCE
3: vPE2 with LowLatency?
DC A1
vPE1
20001
ToR
20002
Spine
20003
WAN
METRO A
LSR
17002
LSR
16003
METRO B
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
Inter-domain SLA with scale and simplicity
No RSVP, no midpoint state, no tunnel to configure !!
BRKRST-3122
38
Anycast SIDs for pairs of border nodes

DC A1
vPE1
20001
ToR
20002
Spine
20003
17901
17901
17901
WAN
METRO A
LSR
17002
16901
16901
16901
LSR
16003
METRO B
16902
16902
16902
LSR
18002
DC B2
Spine
20003
18901
ToR
20002
vPE2
20001
18901
18901
Anycast SIDs provide for better ECMP and High Availability
BRKRST-3122
39
Reachability with Anycast SID
4: {16902, 18901, 20001}
1: V via vPE2
VPN-LABEL: 99999
BGP
RR
2: V via vPE2
VPN-LABEL: 99999
SR
PCE
DC A1
vPE1
20001
ToR
20002
Spine
20003
WAN
METRO A
17901
LSR
17002
16901
LSR
16003
METRO B
16902
LSR
18002
DC B2
18901
Spine
20003
ToR
20002
vPE2
20001
3: vPE2 ?
17901
16901
16902
16902
18901
18901
Better load-balancing: ECMP across border routers
Better availability: sub-50msec upon remote aggregation router failure
Better control plane scalability: no PCE re-computation, no PCEP update, no

FIB update
BRKRST-3122
40
Binding SID to stitch Policies
SR
PCE
3: REPLY {16001, 4001, 20001}

1: REPORT {16003, 16002, 18002, 18001}, UP,
BindingSID 4001
instead of
{16001, 16003, 16002, 18002, 18001, 20001}
2: vPE2 with Min LAT?
DC A1
vPE1
20001
ToR
20002
Spine
20003
METRO A
LSR
17002
WAN
METRO B
LSR
16003
LSR
18002
DC B2
Spine
20003
ToR
20002
vPE2
20001
End-to-end policies can be composed from more basic ones
An SRTE policy is bound by default to a Binding SID

RSVP-TE tunnels can also be bound to a Binding SID and hence RSVP-TE tunnels can
be used within an end-to-end SR policy
Shorter SID list and churn isolation between domains
Even if the WAN-MetroA sub-path changes, the related Binding SID 4001 is constant
BRKRST-3122
41
Seamless Transition
DCI2/32 via AGG2
PREFIX-SID: 18001
2:
DC A1
vPE1
20001
ToR
20002
BGP
RR
vPE2/32 via DCI2

PREFIX-SID: 20001
Spine
20003
DCI2/32 via AGG2

PREFIX-SID: 18001
METRO A
LSR
17002
WAN
LSR
16003
vPE2/32 via DCI2

PREFIX-SID: 20001
1:
METRO B
DC B2
Spine
20003
LSR
18002
ToR
20002
vPE2
20001
3: vPE2 with Low Latency?

4: {16001, 16002, 18001, 20001}
SR
PCE
Best-effort reachability could be provided by BGP3107
ODN and SRTE / PCE provides interdomain reachability with SLA requirements
Eventually, migration of more/all services over SR PCE

BRKRST-3122
42

Inter-Domain Disjoint Services
Example: Two Disjoint Inter-domain PWs

{20003, 16001, 16002,
18001, 20001}
SR
PCE
vPE2 disjoint group 7

vPE22 disjoint group 7
DC A1
vPE1
20001
ToR2
20002
Spine3
20003
vPE11
20011
ToR12
20012
Spine13
20013
METRO A
LSR
17002
{20013, 16011, 16012,

18011, 20021}
WAN
PW1
LSR
16003
METRO B
LSR
18002
DC B2
Spine4
20003
ToR3
20002
vPE2
20001
Spine24
20023
ToR23
20022
vPE22
20021
PW2
ODN/SR-PCE automated compute disjoint paths for PW1 and PW2
PW1 and PW2 do not share the same headend, neither the same tailend

BRKRST-3122
44
Example: Inter-Domain PW - Disjoint Primary / Backup paths

1: Two disjoint paths to vPE2
DC A1
SR
PCE1
Primary
vPE1
20001
2: PRIMARY: {17001, 16001, 16003,

18001, 20001}
SECONDARY: {17011, 16011, 16013,
18011, 20001}
ToR
20002
WAN
METRO A
Spine1
20003
DCI1
17001
17901
Spine2
20004
DCI11
17011
17901
Pri
METRO B
AGG1
16001
16901
LSR
16003
AGG2
16002
16902
AGG11
16011
16901
Sec
AGG12
16012
16902
LSR
17002
LSR
18002
DC B2
DCI2
18001
18901
Spine
20003
DCI11
18011
18901
Spine2
20004
ToR
20002
ODN/SR-PCE automatically computes disjoint primary/sec paths for the PW
sBFD runs at 3x50msec on each SRTE path
Upon failure detection of the primary, the secondary SRTE Path is used
vPE2
20001

BRKRST-3122
45
Topology Independent LFA (TI-LFA)
TI-LFA - Benefits
50msec Protection upon local link, node or SRLG failure
Simple to operate and understand

automatically computed by the routers IGP process (ISIS and OSPF)
100% coverage across any topology
predictable (backup = post convergence)
Optimum backup path
leverages the post-convergence path, planned to carry the traffic

avoid any intermediate flap via alternate path
Incremental deployment
also protects LDP and IP traffic
BRKRST-3122
47
Automated Per-Destination optimization
Demo
2s computes a primary path to 5
PE4
Source
FIB of 2 for destination 5

Incoming Label: 16005
Primary: SWAP 16005 for 16005, oif: 3
100
100
Default metric: 10
BRKRST-3122
8
Dest2
48
Flexible Link vs Node vs SRLG protection
2 checks the protection preference

for the primary interface of the
destination
Demo
PE4
Source
100
100
Link protection (illustration

assumption)
Node protection
SRLG protection
Default metric: 10
BRKRST-3122
8
Dest2
49
Automated and Optimum
2 computes the post-convergence

path if the preferred failure would
occur
Optimality: the operator planned and

dimensioned the post-convergence
path to carry the traffic in the failure
case
Demo
Source
2 uses SR to encode the postconvergence path in a loop-free

manner
2 updates the FIB with the backup
path to 5
PE4
100
100
Default metric: 10
8
Dest2
FIB of 2 for destination 5

Incoming Label: 16005
Primary: SWAP 16005 for 16005, oif: 3
Backup: SWAP 16005 for 16005, PUSH 16007, oif: 6
BRKRST-3122
50
Do we need many SIDs? No!
BRKRST-3122
51
Microloop Avoidance
Microloops are a day-1 IP drawback

Upon link down convergence
IP hop-by-hop routing may induce microloop at any topology transition
Link up/down, metric up/down
4
5
1000
8
Illustration for the post-convergence microloop

impacting traffic from 1 to 9 after link45 going
down. Default link metric 10
BRKRST-3122
Pre-convergence Path
Post-convergence Path
53
SR Microloop Avoidance
Prevent any microloop upon isolated convergence due to
Demo
link up/down event & metric increase/decrease event
2-stage convergence
Stage 1: non-looping SID lists to implement the post-convergence path

Stage 2: post-convergence path
If multiple back-to-back convergences, fall back to native IP convergence
microloop avoidance segment-routing

FIB @ 1 for Destination 9
Initially: {16009} OIF 2
Stage1: {16006, 24065, 16009}
Stage2: {16009} OIF 8
4
5
1000
8
9
Explicit Post-convergence Path
BRKRST-3122
54
Illustration Link Down

Default link metric 10

Initially: OIF to 2
Stage1: {16006, 24065, 16009}
Finally (stage2): OIF 8
Initially: OIF to 1
Stage1: {16006, 24065, 16009}
Initially: OIF to 8
Stage1: {16006, 24065, 16009}
Demo
2
1000
Illustration for the post-convergence
microloop impacting traffic from 1 to 9
after link45 going down

Initially: OIF to 7
Stage1: {24065, 16009}
No microloop can occur thanks to the 2-stage convergence and the use of nonlooping SID lists to implement the post-convergence path in stage1
BRKRST-3122
55
Conclusion
Conclusion
Functionality never seen before
SR is fundamental architecture for modern IP network
Unified Fabric with Policy through DC, Metro and WAN
Simplification through Automation and protocol removal
Strong operator endorsement
Multi vendor consensus
Impressive deployment and velocity
BRKRST-3122
57
Resources
Stay Informed - Tutorials, Conferences, IETF, Open-source SW
http://www.segment-routing.net/
Join us Segment Routing @ LinkedIN
Get in Touch
ask-segment-routing@cisco.com
Latest SR Demonstrations
On-demand Next-Hop and SR PCE

TI-LFA Node protection
Microloop Avoidance
SRv6 Spray use-case
Segment Routing book
Pre-order available now!

BRKRST-3122
58
Complete Your Online Session Evaluation
Give us your feedback to be

entered into a Daily Survey
Drawing. A daily winner will
receive a $750 Amazon gift card.
Complete your session surveys

through the Cisco Live mobile
app or from the Session Catalog
on CiscoLive.com/us.
Dont forget: Cisco Live sessions will be available
for viewing on-demand after the event at
CiscoLive.com/Online
BRKRST-3122
59
Continue Your Education
Demos in the Cisco campus
Walk-in Self-Paced Labs
Lunch & Learn
Meet the Engineer 1:1 meetings
Related sessions
BRKRST-3122
60
Segment Routing opportunities at CiscoLive!
BRKRST-2124: Introduction to Segment Routing
LABSPG-2012: Next Generation Service Provider Network using Segment

Routing & BIER
In the Walk-in Self-Paced (WISP) lab area of the hub until 5pm today!
BRKDCN-2050: Segment Routing in Datacenter using Nexus 9000 and 3000
Presented Monday view session materials on CiscoLive.com
At 1pm today! South Pacific B, Lower Level
LTRMPL-2104: Cisco WAN Automation Engine (WAE) Network

Programmability with Segment Routing
BRKRST-3122
61
Thank you

Segment Routing-Technology &amp; Use Cases

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Segment Routing-Technology &amp; Use Cases

Uploaded by

Copyright:

Available Formats

Segment Routing: Technology

Update and Advanced Use-Cases

Segment Routing Executive Summary

Incremental Deployment Use-Cases

Inter-Domain Policy at Scale

Segment: an identifier for any type of instruction

MPLS: an ordered list of segments is represented as a stack of labels

First deployments in 2015 just 15 months after FCS !!!

Strong start in 2016 with many new deployments

Strong commitment for standardization and

SPRING Working-Group (started Nov 2013)

All key documents are WG-status

Over 25 drafts maintained by SR team

Over 50% are WG status

Several interop reports are available

First RFC document - RFC 7855 (May 2016)

Foundation for modern IP/MPLS networking

Solution to unsolved problems

End-to-end policy, local and/or centralized PCE, 50msec protection, microloop

Set of few, well-chosen building blocks

Granular traffic engineering with minimal network state

SR/LDP interworking, SR/RSVP-TE interworking, ship-in-the-night co-existence

Decoupled data and control planes

Thorough understanding of SR tutorial

You should know

Use-Cases in this presentation described for SR/MPLS

Same concept applies to SRv6 (native IPv6 SR extension header, no MPLS)

Shortest-path to the prefix

Label = 16000 + Index

Forward on the IGP adjacency

Advertised as label value

But only local adjacency SIDs are

All nodes use default SRGB

Steer traffic on any path through the

Path is specified by list of segments

No per-flow state is created

IS-IS, OSPF, BGP all supported

Inter-Domain Policy at Scale

Segment Routing use-case aiming to scale the network to support hundreds of

Applicable to the interconnection of massive-scale DC's and/or large

Principles are equally applicable to a network of any size

SRGB and SID allocation

Homogenous end-to-end SRGB for simplicity

Locally Unique Prefix SIDs for Datacenters

IGP/SR within WAN and Metro Domains

Each domain runs ISIS/OSPF SR

Incremental deployment and seamless interworking with LDP

Segment Routing in the Datacenter

Datacenter fabric runs BGP SR

Example: 20006 is the BGP Prefix SID to DCI6

Inter-Domain Policy at Scale

Intra-Domain Routing DC A1 and DC B2

Intra-Domain Routing Metro A and Metro B

ISIS LSP of AGG2

In a metro, BGP/SR or ISIS-OSPF/SR are likely, both illustrated here

Example: Metro A: BGP/SR

Intra-Domain Routing WAN

ISIS LSP of AGG1

ISIS / OSPF SR in WAN

WAN aggs are re-distributed down to Metro and DC

Nothing is redistributed up !!!

How does vPE1 reaches vPE2?

ISIS LSP of AGG2

Redistribution: from center to leaves

Segment Routing-Technology & Use Cases

Segment Routing-Technology & Use Cases