VNQM is pretty useful in determining the performance metrics between 2 sites. However, it seems that the representation and calculations of 2 key metrics RTT & Latency for IP SLA Operations may be incorrect.
There is a pretty good article that provides how VNQM calculates the RTT & Latency.
From the above article, it is very evident that RTT should be the largest value and Latency should be approx half of the RTT (measured between Source --> Destination or Destination --> Source.
See some IP SLA screenshots from the HCO demo site - the RTT and Latency values are equal both 15 ms. This has been observed not just in the demo website but also in live production environments.
Even more interesting is the other output, where Latency > RTT. Latency is 27 ms while RTT is 5 ms.
We have registered a Case with Support (where real production environment observations have been shared). Case# 01890893
Has anyone encountered such issues before? Any pointers on how we could address this?