SD-WAN: Cloud OnRamp for software-as-a-service

Software-as-a-service applications reside on the Internet, and to achieve the best experience for the user, the best performing internet exit point needs to be used.

When Cloud OnRamp is configured for a software-as-a-service application within vManage; the SD-WAN router at remote sites will probe the software-as-a-service application with small HTTP requests through each exit SD-Router to measure latency and loss.

Based on the results returned to each SD-Router, Cloud OnRamp will know which router will be best placed to be used for the software-as-a-service application.

If there is a change in performance from the continuously HTTP probes sent from each SD-WAN router, the exit point can be easily changed again.

Cloud OnRamp will also monitor sessions between SD-WAN sites and hubs using a version of Bidirectional Forwarding Detection (BFD). Bidirectional Forwarding Detection is normally used for fast detection of path failure between two routers, but in this case it can be used to measure performance of loss, latency, and jitter between two SD-WAN routers.

Not only does the performance between each SD-WAN router and the software-as-a-service need to be measured; but the performance between each SD-WAN router needs to be measured too to ensure that re-routing traffic for a software-as-a-service site via an alternative hub isn’t hampered by the internal transmission of traffic.

The reason BFD is not used over an internet access circuit, is there is no other SD-WAN router to be queried on the other side of the circuit, unlike within the SD-WAN fabric.

The SD-WAN regional hub router will report it’s statistics to other routers in a Overlay Management Protocol message to the vSmart Controllers.

These statistics appear to the administrator as part of a Viptela Quality of Experience score, vQOE. 10 is the best quality connection and 0 is the worst.



, ,




Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.