Announcement

Collapse
No announcement yet.
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Virtual SIP Trunk Switch (VMware)

    Is anyone out there using the virtual ShoreTel trunk switch? We have a couple running in production in our VMware cluster. ShoreTel's recommendation of 24 vCPUs is completely over provisioned which doesn't run the most optimal in VMWare. Also we found that the new Diagnostic & Monitoring has some bugs in vSIP trunks. Just wondering if anyone else has any experience with these?
    Last edited by LDS; 12-08-2016, 10:42 AM. Reason: typo

  • #2
    We've implemented a few of them. I think isn't the recommendation 1 core for every 50 media streams. And there are 2 streams per call, so in other words 1 core for every 25 trunks using G.711, less streams for 724. Anyhow it does seem a little high when you are getting up over 500 trunks.

    I would say though, although it's pretty hefty requirement, if you don't meet the requirement then that will instantly become the blame of any problems.

    As for D&M, i'ts buggy a lot of places not just sip trunks.


    Comment


    • #3
      I have not had a chance to install virtual SIP trunks.
      Looking at the install guide, 24 cores supports up to 500 SIP trunks and 2000 audio streams
      - .Every core can handle 80 to 100 streams in G.711 signaling.
      - . Every core can handle 20 to 50 streams in G.729 signaling.
      On the low end, 2 cores supports 50 SIP trunks.

      I am assuming that they used this guide when suggesting the number of cores to provision. My guess is that the guide is designed for better safe then sorry.

      Comment


      • #4
        I should revisit the install guide. We do have 400 trunks but it seems like even under moderate to high load it's using less than 30% of the CPU resources available to it.
        VMware suggests over provisioning vCPUs will actually hurt performance if they aren't fully utilized. I understand ShoreTel probably gives a general recommendation, and we might consider trying to "Right Size" the VM according to our environment.

        It would be really great if they could fix D&M though. Initially my partner support tried to tell me this was not a bug!! The Trunks Occupancy % is clearly wrong. Our call center is closed at night yet this shows %50 usage all the time vsip-dm.PNG
        Attached Files

        Comment


        • #5
          I don't think it has to do with actual "load", I think it has to do more with the number of threads it can process simultaneously.

          Comment

          Working...
          X