The invocation service allows you to execute an agent on one or more cluster members. Unlike entry processors, which are always a target for execution against specific cache entries, invocation service agents are targeted for execution on one or more cluster members. This makes them more suitable for cluster-wide management tasks than for the actual processing of data.
Another difference is that that the invocation service agents can be executed both synchronously and asynchronously, while the entry processors are always executed synchronously.
Unfortunately, the invocation service has one significant downside as well—unlike entry processors, which automatically fail-over to the backup node if the primary node fails and provide ...