You are here: Pexip Infinity overview > Distributed architecture > Load balancing, redundancy and scalability

Load balancing, redundancy and scalability

This topic describes the load balancing, redundancy and scalability aspects of Pexip Infinity.

Load balancing

Pexip Infinity load-balances intelligently across all Conferencing Nodes that are grouped within a location (Platform configuration > Locations), according to the following logic:

  • Signaling is always handled by the node that originally receives the call. Signaling cannot be moved, and always remains routed to the original node in the local location.
  • If a node in that location is already handling media for this conference, and it has spare capacity, then that node will also handle the media on the new call.
  • If no nodes are handling media for this conference, or all nodes currently handling media for this conference have no spare capacity, then media processing will occur on the least loaded node in the location. (Note that no more than three nodes per location can handle the media for a particular conference instance.)
  • If more than one node is now involved in media processing, a backplane is created between the nodes.
  • If a location has no spare capacity for this conference, the least loaded node in the overflow location is used, if configured.

For more information see Handling of media and signaling.

Load balancing of signaling between Conferencing Nodes in a single location is achieved via your call management system. You should configure it so that calls from each configured set of endpoints are routed to a configured set of Conferencing Nodes. You should configure DNS SRV records to identify all Conferencing Nodes in a location.

For further information about how to configure your specific call management system to work with Pexip Infinity, see the following documentation:


Management Node

The Management Node is used to create Conferencing Nodes, configure conference settings, manage licenses and collate system logs. It does not handle any call processing, so if it were to become unavailable the Pexip Infinity service would be unaffected. However:

  • you would not be able to make changes to services (Virtual Meeting Rooms, Virtual Receptions and so on)
  • logging would be affected: the Management Node would not receive and collate logs from the Conferencing Nodes. This can be mitigated by Using a syslog server to collate logs
  • after 14 days, licenses would cease to be allocated and the Pexip Infinity service would no longer allow calls.

Conferencing Node

You should ensure that your call control system is set up so that, in the event of a Conferencing Node failing, new calls to the conference are routed to another available Conferencing Node.

If multiple Conferencing Nodes are configured in a single location, new calls to a conference can be routed by the call control to only those nodes that are currently available. To achieve this, you must configure the call control system with all possible Conferencing Nodes (either explicitly or through DNS records), and for liveness checks to be carried out.

Note that there is no live fail-over; calls in progress on a Conferencing Node that fails will be lost, and participants will need to redial the Virtual Meeting Room or Virtual Auditorium alias in order to reconnect to the conference.


To increase the capacity of the Pexip Infinity platform, simply deploy one or more new Conferencing Nodes from the Management Node.

You may also need to increase the number of licenses.