Basically I am trying to get my head around the differences and advantages of a cluster over a cell. I am currently working through the WAS v5 redbook but this is slow progress and a down to earth explanation could help me out heaps.
To use a simple deployment scenario as an example, if you have 2 IHS boxes splitting requests across a cell of 2 WAS instances (each one on a different physical box), what is the advantage of a cluster over this setup?
I apologise for the diagram but was trying to illustrate the 2 IHS servers splitting requests across the 2 WAS boxes.
From what I understand, they are in a cell and managed by the Deployment manager. So if I install a new application, the deployment manager will install it across each node in the cell. So is it right that each node in a cell is a copy of each other?
Also, in the above setup, if WAS01 goes down, then both IHS boxes will not route requests to it and all client interaction will be to WAS02. So if you replicated this environment except had WAS01 and WAS02 in a cluster, what is the difference/advantage? I am aware that each member of a cluster must be identical (clones) but isn't each node in a cell identical as well due to the fact that applications are installed across each node in thhe cell?
Thanks in advance for your help.
[ February 01, 2005: Message edited by: Sean McGaughey ] [ February 01, 2005: Message edited by: Sean McGaughey ]