COTS middleware for real-time distributed interactive simulations - fact or fiction?

Abstract

Current and planned real-time distributed interactive simulation systems must collaborate with many event sources, provide interactive response to human operators, and respond jexibly to unanticipated factors arising at run-time. Moreovel; large-scale federated simulation systems must run largely autonomously, shielding operators from unnecessary details, while simultaneously communicating and responding to events at an accelerated operational tempo. In next-generation simulation “systems of systems” it will be hard to predict or even approximate system conjigurations or workload mixes a priori. While it is possible in theory to develop complex real-time distributed interactive simulation systems from scratch, contemporary economic and organizational constraints, as well as increasingly complex requirements, are making it implausible to do so in practice. Thus, the role played by “commodity-off-the-shelf” (COTS) hardware and software has become increasingly significant. In its formative years, COTS middleware, such as CORBA, was generally unsuited for real-time distributed interactive simulation systems due to its lack of I. Standard QoS inte~aces in standard specifications 2. QoS enforcement mechanisms 3. Real-time programming features and 4. Optimizations for eficiency, predictability, and dependability. Howevel; the current generation of middleware specifications and implementations provide much better QoS support for real-time distributed interactive simulation systems. This talk describes how the evolution of middleware capabilities, particularly the OMG Real-time CORBA standardization effort and the DMSO HLMRTI, and the maturation of COTS middleware implementations make it increasingly well-suited as the infrastructure for real-time distributed interactive simulation systems. The material in this talk is based on experience gained with the TA0 Real-time CORBA ORB and the SAIC implementation of the HMRTI, which is based on TAO.