Typical latency to expect between Azure VMs of the same Proximity Placement Group

New Contributor


I was looking at measuring and improving latency between VMs of the same PPG. I wonder what the values are that could be expected, in order to know whether the values I am seeing are "normal" or "bad".

I need low latency between Windows and Linux VMs, so the steps provided here: Test Azure virtual machine network latency in an Azure virtual network | Microsoft Learn don't help to measure exact latency unfortunately. However, I used hrPing to determine the values outside of working hours and am still getting 1.3ms on average.

I know VM size affects the results, so my sizes for reference:

  • Linux: Standard E48ds v4 (48 vcpus, 384 GiB memory), w/accelerated networking
  • Windows: Standard B12ms (12 vcpus, 48 GiB memory), w/accelerated networking

Is that a reasonable result to expect?

Kindly let me know your thoughts.

Thank you!


5 Replies

Hello @Chandrasekhar_Arya ,

As I described, accelerated networking is already enabled. 



So since we are setting up a customer on Huawei Cloud currently, I decided to test latency there between two (much smaller) virtual machines. The average ping came back at just 0.2ms, so by order of magnitudes better than what we're seeing on Azure.
Still waiting on anybody to share comparable numbers, is that difficult to get?
Thanks and best regards,
best response confirmed by Joerg_Aldinger (New Contributor)

@Joerg_Aldinger I have obtained similar results in my tests (1.4ms). I have collected my results in this blog post 



@nicold72 Thank you very much for your detailed blog post. I couldn't find anything similar anywhere. I will make sure to bookmark it for future reference.

It also confirms my suspicion, that my response times are completely normal.

Thanks again and my best regards to you!