Storage Area Network: Architecture

Clustering

Clustering is the ability to combine multiple systems in such a way that they provide services a single system could not. Clustering is used for fault tolerance, parallel processing, and load balancing.

High Availability

High availability clustering ensures that critical business applications are online when needed by removing the physical server as a single point-of-failure. The systems that are clustered together are interconnected via a private LAN and utilize this LAN for a heartbeat mechanism to determine the health of each system. In the event that a hardware or software failure occurs in any system (node), the applications currently running on that node are then migrated to a surviving node.

The effort required to restart the application varies greatly depending on the type of application and the amount of data associated with the application. Ideally, when an application switches from one node to another, the user experiences no disruption at all and isn't even aware that the job has been switched to a different server.

High Performance Computing Cluster

High Performance Computing Cluster (HPCC) combines multiple Symmetric Multi-Processor (SMP) computer systems together with high-speed interconnects to achieve the raw-computing power of classic supercomputers. These clusters work in tandem to complete a single request by dividing the work among the server nodes, reassembling the results and presenting them to the client as if a single system did the work.

Load Balancing

Server farms that can distribute requests to the same application among multiple independent servers are referred to as "load balancing". The term load balancing applies to clusters that include some number of nodes processing requests for the same type of application, often Web servers, streaming media servers, terminal servers, or read-only FTP and file servers. There are software-based mechanisms to perform load balancing, such as Microsoft's Network Load Balancing (NLB) and Red Hat's HA 1.0 (Linux) as well as appliance-based mechanisms, such as the PowerApp^TM BIG-IP load balancing appliance server.