How to protect MSCS

Microsoft’s Cluster Service (MSCS) has brought great advances for the availability of Windows NT based applications by providing a standard platform to develop “highly available” or “cluster aware” applications that can be moved or restarted on an alternate server in the event of certain failure conditions.

The Problem - Shared Storage / Geographic Limitations

All cluster nodes must be able to access the data required by the clustered applications. MSCS provides this capability by switching control of a shared storage subsystem between nodes if you have a certified configuration of servers, disk controllers and disk subsystems. The obvious downside of this configuration is that it provides only a single copy of the data, in a single location leaving the cluster exposed to a single point of failure in the disk subsystem or storage bus. This also restricts the location of cluster nodes and their storage to within several feet if connected by SCSI, and to a maximum of several kilometers if connected by Fibre Channel. This shared storage is a single point of failure making the entire cluster vulnerable to storage related or geographic disasters.

Running Microsoft Exchange or SQL on a Cluster?

Unfortunately, MSCS has some inherent architectural limitations, such as a shared single storage solution and node/distance limitations.

• All MSCS nodes must share the same physical storage. If the storage technology were to fail, none of the nodes would be able to function. This means that the shared data solution becomes a single point of failure.
• Because the nodes are assumed to share the solution, the existing MSCS technology does not provide for the nodes to be geographically separated. This means that if a single building (or even just the computer room) were to be impacted, the entire solution would be affected.

To address these limitations, NSI's replication technologies can be used to significantly increase the availability of Microsoft Exchange.

The Solution - Real-Time Replication

Double-Take protects clustered MSCS data using real-time data replication to provide multiple copies of application data from both nodes of the cluster on a standby server or cluster - eliminating any single point of failure. Data is synchronized using standard TCP/IP connectivity allowing the target replica to be located almost anywhere.

Please contact one of our technical presales staff if you have any questions.