Sep 02, 2014

Voice resiliency is a great feature of Lync Server 2013 which ensures the voice functionality (ability to make and receive calls) is available to users in the event of data centre or a pool level failure. Lync Server 2010 had only PSTN resilience, but Lync Server 2013 offers resiliency to various Lync workloads. Before we start discussing the various resiliency features, let us quickly go through the client registration process in Lync Server 2013. The most common method of offering voice resiliency is by creating multiple Lync Front End pools. Each site contains a Front End pool, which is paired with a corresponding Front End pool from another site. Both sites are active and are assigned different sets of users. When the primary pool fails or is not available, the backup pool now supports both sets of users. The Lync Server Backup Service provides real-time data replication to keep the pools synchronised. The Backup Service is a new feature in Lync Server 2013 that is installed on a Front End pool when you pair it with another Front End pool. There is no restriction on the distance between two sets of data, but it is recommend that you use two data centres in the same world region, with high-speed links between them. A Mediation Server is a necessary component for implementing Enterprise Voice and dial-in conferencing in Lync 2013. The Mediation Server translates signalling and, in some configurations, media between your internal Lync Server infrastructure and a public switched telephone network (PSTN) gateway, IP-PBX, or a Session Initiation Protocol (SIP) trunk. It is crucial that you have a proper resiliency plan for the mediation pool. An incoming call first reaches the Mediation Server and then has to be handed off to the front-end pool which is the next hop. If this next hop is unavailable, an alternative next-hop will be selected. The Mediation Server tracks its next-hop proxy and the backup next-hop proxy by sending out periodic option polls. The backup next-hop proxy is defined by Pool Pairing (associated backup pool), which is defined in the Topology Builder. If the primary next-hop proxy is unavailable (failure to answer to five option polls in a row), new invites from gateways are sent to the backup next-hop proxy. Additionally, a 10-second timer is used for incoming calls, so if the primary next-hop proxy is used for a call and no SIP response is received within this time, the call is rerouted to the backup next-hop proxy. To improve branch resiliency, you have a number of options available;

1. Install redundant WAN connections.
2. Deploy a Lync Pool on each branch site, but this will incur high additional infrastructure cost, licenses, and server hardware. It also adds additional administrative overhead as it requires additional IT staff on each branch site.
3. Deploy PSTN connectivity on each branch site, however this requires a Lync Pool and Mediation Server on each site to function in case of WAN failure.
4. The most common method of implementing branch site resiliency is to implement a survivable Branch Appliances or a survivable branch server.

These purpose-built appliances are optimised to provide resilient communication to branch offices and thus, enhancing user productivity.