Thanks to all who joined us for the Cisco DemoFriday where we learned how to optimize your network, offer better SLA and control your capacity planning operation cost. Read the full Multi-Domain Network Optimization and Simplification Q&A from the event below:
Multi-Domain Network Optimization and Simplification – Why and How… Q&A
Thanks to all who joined us for the Cisco DemoFriday, Multi-Domain Network Optimization and Simplification – Why and How. In the webinar, SDxCentral and Cisco reviewed what the key drivers are to enable multi-domain WAN optimization and simplification, integration between XTC (XR Transport Controller), WAE and NSO, and how SDN is able to be deployed in multi-vendor, multi-domain and large scale environments.
After the webinar, we took questions from audience, which you can read below.
What other type of policy optimizations or constrains have been implemented in XTC?
Cisco: Today, XTC is able to compute paths based on IGP, TE metric and even Disjointness. The plan is to continue to augment its capabilities and add further constrains such as Performance metrics.
Can this solution work in a multi-vendor environment?
Cisco: Yes, this solution has been designed to be multivendor. The several components communicate with each other using standard protocols and interfaces; such as BGP-LS, PCEP, Netconf. And lastly, segment routing is also a standards-based solution.
Can I run XTC on any router running IOS-XR?
Cisco: Correct. The PCE server functionality as well as the native SR-TE algorithms are part of the IOS-XR image. So, one can use a virtual or physical XR router for this function. Obviously, the key requirement to the PCE is to be able to learn the entire topology.
How quickly can the controller (or the solution) react to network failures?
Cisco: Good question, this is where the balance of distributed intelligence and centralized optimization comes to the picture. With Segment Routing, the network is self-healing and able to protect against node and link failures in 50 msec. But obviously, the controller and its dynamic BGP-LS feed will learn of the topology changes and be able to react to the failure by potentially triggering re-optimizations.
Would the solution allow an application to directly communicate with the controller. I would like to to provide it an already-computed path?
Cisco: Yes indeed, XTC north-bound API allows for this. So that the controller can be used to “simply” deploy the SR-TE policies.