Udostępnij za pośrednictwem


The conferencing user model in Lync Server 2013

 

Topic Last Modified: 2012-10-22

A critical part of the Lync Server conferencing user model is meeting size. After collecting data from the multiple data points (as described in the previous section), we determined the following:

  • Most meetings are actually small collaborative meetings with an average of four to six participants

  • Approximately 80 percent of meetings have fewer than 20 participants.

  • 99.98 percent of meetings have fewer than 100 participants.

In addition to meeting size, the conferencing user model also takes into account a variety of factors, such as:

  • Concurrent meetings   How many users are expected to be in meetings at the same time?

  • Media mix   What types of media are available and expected to be used by users in meetings?

  • User types   Are users internal users, remote users, federated users, or anonymous users?

  • Meeting ramp up time   How long does it take for all users of a meeting to join a meeting?

For details about the user model, see User models in Lync Server 2013.

To determine the number of meetings and users to use for testing, we did the following:

  • Took the total number of users in an organization (for example, 80,000 users) and multiplied it by the meeting concurrency rate (for example, 5% of all users) to determine the total number of users expected to be in meetings at the same time (in this example, 4000 users).

  • Divided the total number of users by the number of Lync Server 2013, Front End Servers in the deployment (for example, 8 servers) to determine the estimated number of meeting participants per Front End Server (in this example, 500 users per Front End Server).

  • Divided the number of users per Front End Server by the average meeting size (for example, 4 users) to determine the estimated average number of meetings per Front End Server (in this example, 125 meetings per Front End Server).

  • To get the per media load on each Front End Server, we estimated the media mix. For example, assuming that 75% of the meetings require more than just audio support and 50% of those meetings require application sharing, an average of 47 meetings and 188 users connect concurrently to each Front End Server for application sharing.

  • Tested a variety of meeting sizes (based our user model of up to 250 users in a shared pool) to ensure server scalability.