Clustering servers for high availability and failover

You can provide failover and load balancing in your deployment by creating a cluster of multiple IBM® Sametime® servers of the same type. All Sametime servers can be clustered except for the Sametime System Console, and the Video MCU and TURN Server components of the Sametime Media Manager.

About this task

Sametime Community Servers are "clustered" by means of the IBM Domino® replication feature, which automatically copies data to each server within the cluster to provide failover. To improve performance, client connections to the cluster can be routed through one or more stand-alone Community Mux servers, one or more Sametime Proxy Servers, or even a rotating DNS server.

The Sametime Video Manager is hosted on IBM WebSphere® Application Server, but is clustered using special procedures to replication the server's data.

Other Sametime servers that are hosted on WebSphere Application Server are clustered using WebSphere Network Deployment. When you cluster these servers, you should deploy two or more WebSphere SIP proxy servers in front of each cluster to provide failover in case one of the product servers fails. Deploying a load balancer in front of the SIP proxy servers lets you distribute workload among the cluster members automatically to improve performance.

Important: For best results, configure all of the members of a WebSphere-based cluster within a single subnet. You can use a network interface card (NIC) on a physical computer, or a logical partition (LPAR); both can be managed through a domain name system (DNS) server that is deployed in the same network configuration.

For information about Load Balancer configurations for new installations, see the Tech Note Sametime: Load balancer configurations for new installations.