In WebSphere world, there are two variations of the product.
(a)
Base WebSphere: This version only runs as a standalone application server. There is no concept of node, nodeagent etc under this product.
(b)
WebSphere ND: ND stands for Network Deployment and this is the version, which allows clustering of application servers.
App Server: This is the process, which hosts the applications. Each app servers runs in its own JVM. Each app server has a process id.
Node: Logical group of managed App Servers. Generally a node represents one physical machine. But one can create multiple nodes on the same physical machine.
Node Agent:Node agents are administrative agents that route administrative requests to app servers
Cell: Cells are arbitrary, logical groupings of one or more nodes in a WebSphere Application Server distributed network.
So to simply one cell can have one or more nodes, which can have one or more app servers.
The process that manages Cell is called Deployment Manager and the process that manages Nodes is called Node Agents.
With respect to the Data Source...
you should consider assigning the Cluster as the scope of the data source....
Hope this helps...