Topology Options and Scaling

Several topology options are available with Persistent Chat, and these are dependent on the overall Lync deployment. Similar to the Lync Front End Services, an instance of Persistent Chat is referred to as a pool, even if just a single server is used. As with other Lync server roles, both front-end and back-end components are included in a Persistent Chat pool. The front-end components include the Persistent Chat service and the Compliance service. Back-end databases for Persistent Chat include the Persistent Chat Store and the Persistent Chat Compliance Store.

With a Standard Edition deployment, the Persistent Chat Server role can be collocated with the Front End Server. The single-server deployment can support up to 20,000 users. However, since all Lync services are hosted using a single system with this topology, Persistent Chat performance will be entirely dependent on the resource usage requirements for all Lync services. With an Enterprise Edition deployment, Persistent Chat cannot be collocated with the Lync Front End Services, and therefore dedicated systems are required. However, the Persistent Chat databases can be collocated on the same SQL Server, or even the same SQL instance, as the Lync Back End databases. Hosting the Persistent Chat Stores on a dedicated SQL Server is of course also an option, and is recommended particularly for larger Lync implementations to ensure the best performance.

Persistent Chat also includes support for multiple-server topologies, with several high-availability and disaster recovery options. Up to eight Persistent Chat Servers can be installed into a single pool, and these servers can be installed across multiple locations for site resiliency. Up to four Persistent Chat Servers in a pool can be active at any time, and each active server can support up to 20,000 concurrent connected endpoints, for a total of 80,000 users. The load is automatically distributed evenly across the active servers, and in the event of a server failure, users are automatically transferred to a remaining active server. When multiple Persistent Chat Servers are used, the file stores and databases are shared among the servers, and the servers freely communicate with each other as needed to form a cohesive Persistent Chat system. The end result is that all chat history is available to any of the servers in the pool, and users connected to different Persistent Chat Servers can freely chat with each other.

Intra-site high-availability can be achieved with the installation of multiple SQL Servers in a datacenter and using SQL mirroring to replicate the data between them. For cross-site disaster recovery, at least one set of Persistent Chat Front End Servers and dedicated Back End database servers must be installed at each location. SQL log shipping is then used to replicate Persistent Chat database information between the two locations. For more detailed information regarding high-availability and disaster recovery options with Persistent Chat, see Chapter 15, “High-Availability and Disaster Recovery.”