Where required a web service could read from a database, or use session state by way of cookies or WS-Session.
When clients talk to the web service they include the identifier of the specific resource that should be used inside the request, encapsulated within the WS-Addressing endpoint reference.
This may be a simple URI address, or it may be complex XML content that helps identify or even fully describe the specific resource in question.
These can be used to read and perhaps write resource state, in a manner somewhat similar to having member variables of an object alongside its methods.
Most fundamental is architectural: are distributed objects with state and operations the best way to represent remote resources?
There are, however, implementations of WSRF that support persistence, clustering and high-availability of resources (for example, in WebSphere Application Server).
Its limited compatibility with the mainstream WS-I architecture created dissent from the UK grid community.