Guest

Cisco TelePresence MCU 4200 Series

Configuring load balancing on MCUs

Configuring load balancing on Cisco TelePresence MCUs

This article applies to the following products:

  • Cisco TelePresence MCU 4200 Series
  • Cisco TelePresence MCU 4500 Series
  • Cisco TelePresence MCU 5300 Series
  • Cisco TelePresence MCU MSE Series

Configuring the MCUs

To load balance conferences between MCUs you need an H.323 gatekeeper that supports load balancing across prefixes, such as Cisco TelePresence Video Communication Server (VCS), GNU GK, or the MCU's built-in gatekeeper. The Cisco TelePresence 4200 Series, 4500 Series, 5300 Series, and MSE Series (all called MCUs below) can be configured so that they share a common prefix. Users dial a number consisting of that prefix (to reach the MCUs) and a suffix (to identify the conference to join). Every call dialing that suffix connects to the same conference, any call dialing a new suffix creates a new conference. The conferences can be load-shared over the MCUs, with each new conference appearing on another MCU.

Do the following on each MCU:

  1. Go to Settings > Gatekeeper and for H.323 gatekeeper usage select Enabled.
  2. Enter the address of the gatekeeper for your system. This can be the built-in gatekeeper or a third-party gatekeeper.
  3. Enter a unique H.323 ID to register for the MCU.
  4. For MCU service prefix enter the shared prefix. Give each MCU the same prefix.
  5. If required, set the Send resource availability indications (see the explanation below).
  6. Click Apply changes.
  7. Go to Settings > Conferences.
  8. For Incoming calls to unknown E.164 number select Create new ad hoc conference.
  9. Select Register ad hoc conferences with gatekeeper.
  10. Click Apply changes.

The load-sharing behavior of gatekeepers varies; if you are using a third-party gatekeeper, please consult the documentation for details. To load balance conferences evenly requires a gatekeeper which supports resource availability indications. The built-in gatekeeper does not currently support them; however, it load-shares conferences in a round-robin mode over the available MCUs. If your configured prefix is 11, then the first call to 11001 creates a conference on the first MCU. Subsequent calls to 11001 are routed to the same MCU and join that conference. A call to 11002 creates a new conference on another MCU. The conferences continue until the last participant leaves.

Send resource availability indicators

To ensure that conferences are routed to blades that can run the conference, set the resource availability indicators in the Settings > Gatekeeper page. These control if and when messages are sent to the gatekeeper indicating that the blade is almost full based on the number of conferences present or number of ports in use. For example, if a blade should only have 4 conferences, then set the threshold for conferences to 4; when the fourth conference begins the blade informs the gatekeeper that no new conferences should be directed to it. New conferences will be sent to other blades; calls to existing conferences on that blade will still be sent to it.

Resource availability indicators are particularly useful for gatekeepers that do not load share conferences over the available blades, but direct all conferences to a single blade. When that blade is full it informs the gatekeeper which can start to fill the next blade. When conferences finish and people disconnect, the blade sends an indication that it can accept new conferences again.

Further reading


December 20th, 2012 TAA_KB_93