ICCA cluster traffic latency.

Unanswered Question
Dec 20th, 2007

We have a global voip implementation being initiated within our organization. In our asain pacific region we would like to cluster our primary and subscriber in Hong Kong with a secondary subscriber in Tokyo. There are latency issues though, working with two of the largest providers in the area (singtel and anc.) They have indicated to us that the best latency that they are currently seeing between data-centers in HK and Tokyo is 55ms. The documentation indicates that the highest possible RTT for ICCA traffic is 40ms. Will it be possible to operate at 60ms? What difficulties will the system have?

Any suggestions?

Appreciated.

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Loading.
Chris Deren Thu, 12/20/2007 - 10:34

You will probably be fine, however Cisco TAC may refuse support once you run into issues. The most critical issue you can expereince is database replication. I would suggest designing a deployment that meets all Cisco guidlines to avaoid issues during operation.

So perhaps splitting cluster accross WAN may be out of the picture :-(

Chris

lewisx_2 Thu, 12/20/2007 - 12:39

I appreciate the reply.

Has anyone run this in an active enviroinment?

I was aware that there could be replication issues, but I am really looking for real world predictability. Anyone currently using latency greater than 40ms intracluster?

mciarfello Sun, 12/23/2007 - 22:47

Put your intercluster stuff for that 55ms site on hold for a little while and check the release notes of upcoming CCM releases.

Don't chance the 40ms timings. You may/will get weird stuff happening.

kolkatagtc Tue, 12/25/2007 - 20:44

Hi,

I have came across this scenario but in my case the latency was more than 60 it was some around 64-65ms. The phones used to show as un-registered alongwith the gateways whereas they are originally registered with the local subscriber.

In SQL server replication used to take longer time, and many of the times used to show that connection could not be established which lead to great problem for exact troubleshooting.

These issues were not there, when we increased our BW. I was using CCM4.1.3 sr5b.

Thanks

lewisx_2 Wed, 12/26/2007 - 13:45

So in your case the latency wasn't the issue then, it was the Bandwidth? After increasing the bandwidth did you have a change or improvement with your latency?

Actions

This Discussion