Monday, February 21, 2011

WebSphere Concepts - Groups and Clusters

It is sometimes confusing about all the WebSphere groups - node group, default group, core group, cluster, and dymanic cluster. Here is a quick summary.

node group -  a group of nodes of similarity within one cell. For example, a cell can have a two node groups of zOS and one node group of nodes on distributed platform. To me, the node group provides logical grouping of application servers and made system administration easier.

default group - usually, on WAS ND, if you do not specify node groups, all in one cell belongs to a defaultnodegroup.

cluster - A cluster is a collection of servers managed together. Clusters behave as a single application server entity to accomplish a job by parallel processing. The intent of clustering is high availability and failover. A frequently ignored important aspect of clustering is that it can be effectively used to increase system resources to a set of applications in order to improve throughput and overall JVM level capacity.
  • all cluster member must be within the same cell
  • all cluster member must be in the same node group
  • a cluster member can only belong to one cluster
  • a cell can have one ore more clusters
  • all members must run the same set of applications
core group - a group of clusters in a high availablity envrinment. All of the application servers of a cluster included in a core group are automatically members of that core group. Core groups enables Websphere Application Server to provide high availability for applications  You can also configure core groups to communicate with each other using the core group bridge. The core groups can communicate within the same cell or across cells. Core group works with high availability manager and communication bridges. The intent of core group is high availability.

dynamic cluster - A dynamic cluster is a concept and mechanism used by WebSphere XD to manage a dynamic WebSphere environment. It monitors performance and load information and is able to dynamically create and remove cluster members based on the workload. The intent of dynamic cluster is dynamic provisioning of JVM as well as application edition management.

1 comment:

Abdul Wahid said...

I know this comes a bit late , but want to mention that “Clustering” might probably come in handy when the user has no knowledge of the available content or if he is aimless, looking for new information. But, in an Enterprise site search , where users knows exactly know what they are looking for (per say, a retail distributor searching for a product document within a site), the percentage of users clicking on a cluster will be very low. So, the usage of clusters will differ based on search context. http://www.vpswebserver.com/windows-vps/clusters in this article you will get more information about clusters.