This is part six of The Twelve Days of Lync’mas, to see an index of all twelve posts click here.
On the sixth day of Lync’mas my UC Team gave to me: 6 core CPUs
Introduction
Remember when multi-core CPUs first came out? Not so long ago I remember thinking, dual cores on one socket, no way! Then quads…. egads, how much processing power do we need?!? Now with Lync Server 2013, we are being advised to put hex core CPUs in our servers! Now, before you go off saying “bloatware” or “typical Microsoft…” it is important to understand the architecture changes behind these recommendations.
Hardware Recommendations Review
First, let’s review the hardware recommendations for Lync Server 2010:
CPU | One of the following:
|
MEMORY | 16 GB |
DISK | Local storage with at least 72 GB free disk space on a 10,000 RPM disk drive |
NETWORK | 1 network adapter required (2 recommended), each 1 Gbps or higher |
Now, the hardware recommendations for Lync Server 2013:
CPU | 64-bit dual processor, hex-core, 2.26 gigahertz (GHz) or higher |
MEMORY | 32 GB |
DISK | 8 or more 10,000-RPM hard disk drives with at least 72 GB free disk space.Two of the disks should use RAID 1, and six should use RAID 10.-OR-Solid state drives (SSDs) which provide performance similar to 8 10,000-RPM mechanical disk drives |
NETWORK | 1 dual-port network adapter, 1 Gbps or higher (2 recommended, which requires teaming with a single MAC address and single IP address) |
As you can see, there is a recommended increase across all resources, memory, disk and network. Why? The answer, in part, is Lync 2013’s “Brick Model” architecture.
Lync Server Front End Pool Architecture
Lync 2010 Pool |
Lync Server 2010 Front End pool architecture was a centralized design with a SQL Server database backend doing a lot of heavy lifting. Each Lync Server Front End Server was tightly coupled to the backend and in constant communication with the SQL database for presence updates, subscriptions as well as all business logic. This tight coupling between Front End Server and the backend as well as the SQL “chattiness” limited the scalability of the Front End pool to a maximum of 10 servers. |
Lync 2013 Pool |
The Lync Server 2013 Front End pool architecture has moved to a “Brick Model” where each Front End server houses the presence database. The Front End servers perform “lazy writes” to the SQL backend database for rehydration and disaster recovery. Moving responsibility for the presence database from the SQL backend to the Front End server is one factor in the increased hardware recommendations. This has resulted in simpler requirements for the SQL backend, greater scalability of the Front End pool and a much improved user experience for Front End server failover. |
Conclusion
While the recommended hardware specification have increased with Lync Server 2013. Hopefully, I have shown that the benefits of the increased recommendations are worth the increase in hardware. By moving the presence database to the Lync Front End Servers the dependency on the SQL backend is greatly reduced leading to greater scale-out capacity for Front End pools as well as an improved user experience in the event of a SQL backend failure.