How Segment Routing Works
When a packet arrives at the SR ingress node, it’s subjected to policy. If the packet satisfies match conditions for an SR path, the SR ingress node encapsulates the packet in an SR tunnel that traverses an SR path, segment by segment.
Each segment in an SR path terminates at a segment endpoint node. When a packet arrives at a segment endpoint, the endpoint examines the outermost packet label or header to attain the corresponding segment. It then pops the outermost label or header and forwards the packet to the next segment endpoint. This process continues until the packet arrives at the final segment endpoint, which may be the SR egress node.
When a packet arrives at the SR egress node, that node determines whether the packet is at the end of its path. If it is, the node removes the SR header information and forwards the packet based on its destination IP address.
Because the transit routers simply forward the packets based on the SR segment identifier (SID), SR can be used to map packets associated with an end user or application to specific network function services. It does this by mapping a path to where the service will be applied, and providing instructions about the service and additional path information from the service gateway to the SR domain egress router.
Segment Routing Benefits
SR’s primary benefit is its ability to simplify the network and reduce resource utilization, making it easier to manage and operate your network.
Other benefits make SR desirable in a network.
- SR reduces the number of nodes that need to be touched for path provisioning and changes. This action enables SR to be more responsive to network changes, making it more agile and flexible than other traffic-engineering solutions.
- SR traffic engineering provides application QoS, and maps network services to end users and applications as they traverse across the network.
- SR provides resiliency through headend restoration and topology-independent loop-free alternate (TI-LFA) technology, which helps with path reliability during network outages.
When used with a WAN PCE Controller, SR offers additional benefits.
- It provides bandwidth reservation with simplified traffic engineering, as the controller has the ability to assign link attributes and path constraints and perform Constrained Shortest Path First (CSPF) calculations.
- It reduces the risk of transition whether you’re deploying SRv6, SR-MPLS, or SRm6 in your network. It does this by providing heterogeneous support for multiple forwarding planes, including MPLS and IPv6.
- It facilitates closed-loop automation by continuously assessing real-time network conditions such as packet flows to network services, monitoring network behavior and performance, and making changes once that can be automatically distributed across the network without having to push the changes to multiple routers via CLI.
- It enhances the end-user quality of experience by defining specific network paths per set requirements (also used for network slicing).
Traffic Engineering with Segment Routing Controllers
An SR controller is a type of SDN controller that provides centralized path computation, traffic engineering, and granular visibility and traffic flow control for SR forwarding planes in service provider and enterprise networks. The SR controller enables network operators to optimize their network infrastructure through proactive monitoring and planning, and by dynamically routing large traffic loads based on specified constraints.
A key benefit of the controllers is their ability to provide bandwidth reservation, which segment routing by itself doesn’t do well. By moving path state information from transit routers into the packet, segment routing eliminates the need for protocols such as Label Distribution Protocol (LDP) and Resource Reservation Protocol-Traffic Engineering (RSVP-TE), which distribute path information across the network. RSVP-TE provides a mechanism for bandwidth reservation, so eliminating it can be problematic in networks where bandwidth reservation is currently implemented.
SR controllers solve this problem because they can monitor the entire network topology and its traffic flows in real time. They use this data to determine the explicit paths that packets should take through the network, and they can also allocate bandwidth to those paths. After the paths are calculated and link bandwidth is allocated, the controller adds this information to its database. The controller factors in existing bandwidth needs as it computes new paths or before it reroutes traffic due to congestion or other dynamic network conditions.
The SR controller performs three basic operations.
Gathering finely granular telemetry data from the network, the controller analyzes the data and optimizes it to assist in taking intelligent actions, such as tunneling traffic through the most efficient network path, meeting SLA requirements, or proactively resolving congestion issues.
NorthStar Controller, Juniper’s WAN software-defined networking (SDN) controller for traffic optimization, provides these capabilities and enables you to view and monitor multiple network domains. NorthStar Controller gives you an end-to-end network view and the ability to achieve interdomain traffic engineering and end-to-end network slicing.
NorthStar Controller also helps with migration to SR by centralizing routing control and giving you the ability to run multiple control planes over the same infrastructure during your migration. The GUI in NorthStar Controller provides a real-time network view with very fine network details that make it simpler to design and operate your network.
SR uses either MPLS (SR-MPLS) or IPv6 (SRv6) on the forwarding plane. Juniper offers both solutions. To overcome SRv6 shortcomings, Juniper has developed SRm6, which reduces the size of SIDs in the SR header using a smaller compressed routing header and mitigates the resource constraint problems seen in wide SRv6 deployments.
The robust, full-featured NorthStar Controller is the only SR controller to provide more than just bandwidth reservation. It accommodates path diversity and can perform CSPF computation, bandwidth calendaring, and other useful functions. A powerful and flexible traffic-engineering solution, NorthStar Controller enables granular visibility and control of SR and IP/MPLS flows in large service provider, cloud, and enterprise networks.