[cisco-voip] CUCM Capacity and Maximum Supported Endpoints

Wes Sisk wsisk at cisco.com
Tue Oct 25 15:31:05 EDT 2011


Reading between the lines it sounds like your question is, "Is this enforced?".  I would respond with:
CSCsc71873    NumDevRegExceeded- Number of registered devices exceeded:Phones cant REG

It is not just a number stored in a database. It is enforced by the ccm process.

Beyond that you said the keyword *megacluster*.  Those deployments get specific instructions for deloyments.

AFAIK the sizing tools are not currently optimized for megacluster deployments:
http://www.cisco.com/cgi-bin/CT/PGWCT/ct.cgi

Your Cisco account still needs to submit megacluster designs to a special process for performance modeling and subsequent approval.  I believe this is the most appropriate answer to your question.

Regards,
Wes

On Oct 25, 2011, at 2:20 PM, Peter Slow wrote:

Hello Boys and Girls,
    I have a question regarding CUCM Capacity. The SRND states that at this time with CUCM 8.x, there is a "Maximum of 40,000 configured and registered Skinny Client Control Protocol (SCCP) or Session Initiation Protocol (SIP) IP phones or SCCP video endpoints per cluster." ...Elaboration on this would be helpful.
* Is the 40K limit on the number of devices configurable in the database, _or_ can we keep adding devices up to a particular point as long as we don't exceed 40K simultaneous registered phone devices?
* What happens if we get to that limit? would CUCM cease to allow new phones to register, or would it stop me with an actual imposed limit of 40K configured phone devices in my database?
* My ultimate goal is to figure out if I should keep growing my megacluster, or if I need to start making plans to turn up a new one.

Background Info:
I am running 8.6.1.21011-1. I have 12 subscribers doing call processing, each of which was built with the 7500 user VMWare OVF template. I am licensed for more. (VMware Installation: 2 vCPU Intel(R) Xeon(R) CPU E5640 @ 2.67GHz etc...)  
I'm at slightly over 16K _registered_ devices today and we have one-to-one redundancy. For each subscriber I've got processing calls, I have one that is almost 100% idle. I have little to no CTI Controlled devices on this cluster.
My current constraining factor is SAN disk space, so i'd like to save as much of that as possible. I have plenty of free UCS processing power and RAM.
I have 4 other similar clusters, all of which need to be able to call each other. New clusters mean a scarer dialing plan but that can be managed.

-Peter Slow
_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip




More information about the cisco-voip mailing list