EXPERT RESPONSE
"Clustering" has become one of the biggest buzz words since SQL Server 2000 and Windows 2000 were introduced due to the improved manageability and reliability. Clustering in SQL Server 2000 is when two or more servers are configured to operate as a single virtual SQL Server to support database transactions. Each node or server in the cluster has the needed files installed locally as well as via a shared disk subsystem to support the SQL Server databases. A heart beat is used between the nodes to ensure each node is functioning properly, if not a failover occurs. Clustering is used to achieve high availability with SQL Server to support 24x7 operations to prevent a hardware failure from stopping an application from servicing client needs.
In a nutshell, with Windows 2000 Advanced Server and Windows 2003, two to 32 nodes can support a cluster to meet business needs. The two primary clustering modes for a two node cluster are:
- ActiveActive – Both nodes in the cluster process database transactions typically at 50% of the total resources because either node can fail-over to the other so that a single node would be responsible for the transactions of both nodes.
- ActivePassive – One node in the cluster processes database transactions and the second node processes database transactions only if the first node fails.
A more recent buzz word in the clustering space is "geo-clusters." Geo-clusters use the same general technology, but the nodes of the cluster are geographically disbursed in different regions, states or even countries. This is an emerging technology, but worth watching as it grows and is adopted by organizations.
For additional information about clustering visit:
For vendor information related to clustering visit:
For More Information
|
