Shared State for Heterogeneous Distributed Systems

InterWeave is a distributed middleware system that supports the sharing of strongly typed, pointerrich data structures across a wide variety of hardware architectures, operating systems, and programming languages. As a complement to RPC/RMI, InterWeave facilitates the rapid development of maintainable code by allowing processes to access shared data using ordinary reads and writes. To maximize performance on bandwidth-limited networks, InterWeave caches data locally, leverages application-specific coherence requirements to minimize the frequency of updates, and employs two-way diffing to update only those data that have actually changed. We have used InterWeave to construct applications in a variety of problem domains, including remote scientific visualization, interactive data mining, and human/computer collaboration. For previously uncached data, InterWeave achieves communication performance comparable to that of parameter passing in traditional RPC systems. For data that have already been cached, relaxed coherence and platform-independent diffs allow InterWeave to significantly outperform the straightforward use of RPC.