3.5. Summary

In this chapter we've seen how to take ordinary data structures and create distributed versions that can be concurrently accessed and manipulated by multiple processes. We started with the most basic example: an individual entry that can act as a shared variable and provide important atomic properties that allow us to safely update its values without explicit locking. These properties follow directly from the key features of JavaSpaces technology. From there we looked at slightly more complicated data structures; while most of these data structures have conventional analogs (arrays) others (like the unordered bag) may have been new to you. There is no limit to the kinds and complexity of distributed data structures you can build using entries: distributed linked lists, hierarchical tree structures, graphs, and so on.

Now that we've gotten the basic building blocks out of the way, we are going to explore their use in two different, but related, domains: synchronization and communication.