Synopsis  
Routing protocols provide a powerful control plane which allows us to define flexible policies on how to handle different types of traffic, as well as pre-compute backup paths to route traffic around failures before the network reconverges. 

However, IP data plane is designed to make forwarding decisions based on destination IP, thereby imposing a serious limitation on routing design. MPLS data plane helps to overcome this limitation by allowing routers to make forwarding decisions based on label value rather than IP address, where label values can encode routing policies, service definitions, QoS treatment etc. MPLS can also provide the underlay for various services  such as L3VPN, L2VPN/EVPN, MVPN, 6PE etc. 

For a couple of decades, MPLS Traffic Engineering (MPLS-TE) has been successfully used to supplement link-state IGP control plane and provide routing based on administrative and bandwidth constraints, as well as quick switchover to pre-computed backup paths, known as Fast Reroute (FRR). MPLS-TE uses IGP (IS-IS or OSPF) to advertise topology, various link attributes such as configured administrative constraints, available bandwidth etc, and Resource Reservation Protocol (RSVP) to signal MPLS labels hop-by-hop, yielding an end-to end Label Switched Path (LSP) which can be used to steer traffic via a path other than the IGP-calculated shortest path. 

The fact that each Label Switch Router (LSR) in MPLS-TE must maintain the state for each LSP it originates, terminates or acts as a transit point for, leads to a serious scalability issue. Various optimizations have been done in an attempt to solve this problem (for example RSVP refresh reduction or hierarchical network designs with targeted LDP over RSVP). They solve the scalability problem to some extent, but increase network complexity. 

Segment Routing reinvents the MPLS data plane for IP routing and eliminates label distribution protocols such as LDP and RSVP. Instead, MPLS labels now become globally significant and are advertised by IGP or BGP together with IP prefixes. Traffic Engineering is also possible, but is stateless for transit nodes. FRR is replaced by Topology-Independent Loop-Free Alternate (TI-LFA) which solves the same problem in a more optimal way.  

Segment Routing emerged as a new technology in 2013 and has made a huge impact with service providers, hyper-scale web providers and large enterprises. As a key enabler for the transformation all networks must go through in the future, it’s no wonder that adoption is growing exponentially. But what will it actually do for your business? 

A. Simplify Network 
Removing protocols and making network operations easier means Segment Routing in your fast track to network simplification. 

B. Make the network more robust 
With Segment Routing, your network is more resilient. Whenever and wherever a node or a link fails in the network, connectivity is restore in under 50 
milliseconds. 

C. Squeeze more out of your network 
Hyperscale cloud providers have learned that by dynamically rerouting traffic, they can reach an overall capacity utilization of 80 percent or more. 

D. Release Innovation 
Segment Routing powers two vital network services, low latency and disjoint Ness 

E. Offer new levels of customer satisfaction  
With an efficient network structure, you have a strong foundation on which to offer a best-in-class end-user experience. 

Objective 
After completing the workshop, participant should be able to : 

• Identify the basic concepts of Segment-Routing MPLS and its differences from MPLS 
• Describe the terminology of Segment, Segment-ID and Segment-Routing 
• Describe how to use the IGP-control-plane (ISIS) to distribute segment 
• Describe how internetworking with Non- SR networks 
• Describe BGP-LU to advertise the segment  
• Describe the use of Transit Fast-Restoration (TI-LFA) and Micro-loop avoidance in SR-MPLS 
• Describe Segment-Routing Traffic Engineering 
• Deploy SR-MPLS using Arista Platform 

Target Audience 
Professionals, Engineers or Technicians that involved in transport provider network to build and manage the operational network. Any other with networking technical background are also welcome to join to enrich their technical knowledge of network 
transport technology  

Pre-requisites 
Before participating in this workshop, each participant should have knowledge of the networking concept: 

• Knowledge of TCP/IP and also its implementation 
• Knowledge of bridging and routing process in layer-2 and layer-3 network 
• Knowledge of Routing protocols  
• Knowledge of MPLS basic concept  

Other Requirements 
So that this workshop can be carried out well, it is recommended that each participant bring their own laptop 

Time Durations  
Two days 

Course Outline 
This course will be conducted in 2 days  
Day-1 for Module 1 & Module 2 
Day-2 for Lab session module 3 

Module 1 : Introduction of Segment-Routing MPLS (SR-MPLS) and introduction of Arista EOS 
• Basic Concept why SR-MPLS  
• Type list of segment, SRGB 
• SRGB in Arista EOS 
• IGP ISIS-SR with node segments 
• Brief overview of SRv6 
• Introducing of Arista EOS 

Module 2 : Segment-Routing MPLS (SR-MPLS) operational 
• Internetworking with non-SR networks 
• SR-MPLS and LDP-MPLS coexistence 
• SR-MPLS internetworking with LDP-MPLS (SR to LDP mapping server, LDP to SR)   
• BGP labelled Unicast (BGP-SR) 
• Transit Fast Restoration in SR ( TI-LFA, Micro-loop avoidance) 
• Segment-Routing Traffic Engineering  

Module 3 : Deploy Segment-Routing MPLS (SR-MPLS)  
• Configure segment types in SR-MPLS 
• Configure ISG ISIS to distribute the segments 
• Configure SR-LDP Coexistence 
• Configure SR internetworking with LDP 
• Configure BGP labelled unicast 
• Configure TI-LFA  
• Configure SR-TE 
• Configure to enable VPN-service (l3vpn over SR-MPLS) 
• Services verification  

Speaker Information

Name : Vidi Ali Ahmad
Title    : System Engineer at Arista Networks