Earlier in the week I replaced a pair of MX100s running 16.16.3 with a pair of MX250s running the same code.
Looking at the graphs below that show the latency and jitter from this site to another local one that is still running a pair of MX100s, you can clearly see when I swapped them out.
The site's links aren't overloaded, with traffic pretty constantly around 100Mbps across a pair of 230/1000 circuits.
I don't think so, as if I look at a different connection (from the same source site) to our main DC, that has MX250s running 16.15, then the overall latency and jitter are now near zero! There are only two graphs this time as the DC MX is in VPN concentrator mode.
Also the MX100s had only been upgraded to 16.16.3 a week or two ago. I'm thinking a fundamental chipset difference...
