Principles and Patterns of High-performance,
Real-time Object Request Brokers

Abstract: Middleware is becoming increasingly important for building flexible communication systems that reduce software development cycle time and effort.  Unfortunately, conventional implementations of middleware like CORBA, DCOM, and Java RMI have historically lacked the efficiency, predictability, and scalability required by real-time and embedded communication systems.  A decade of intensive R&D on design techniques and optimization principle patterns has recently converged, however, to yield high-performance and real-time middleware that can meet end-to-end application Quality of Service (QoS) requirements.

This talk outlines recent advances in Real-time CORBA, focusing on the patterns of design and optimization for (1) QoS-enabled CORBA programming models, (2) ORB architectures that minimize priority inversion and non-determinism, associate client requests with servants in constant time, and implement standard and custom middleware protocols using small memory footprints, (3) CORBA scheduling and events services for adaptive QoS and static/dynamic real-time scheduling, and (4) middleware services that control and manage embedded real-time system components efficiently, scalably, and predictably.

The patterns and performance optimizations covered in the tutorial are based on TAO, which is an open-source real-time ORB that supports end-to-end QoS guarantees over a range of networks and embedded system interconnects.  TAO is currently deployed at Boeing, DARPA Quorum, Lockheed, Lucent, Motorola, Nortel, Raytheon, SAIC, and Siemens, where it is used for real-time avionics, simulations, and telecommunications systems.  Source code, documentation, and technical papers on TAO are available at www.cs.wustl.edu/~schmidt/TAO.html.

Biography: Dr. Schmidt is Director of the Center for Distributed Object Computing and an Associate Professor of Department of Computer Science and the Department of Radiology at Washington University in St. Louis, Missouri, USA.  His research focuses on design patterns, implementation, and experimental analysis of object-oriented frameworks that facilitate the development of high-performance, real-time distributed object computing systems on parallel processing platforms running over high-speed networks and embedded system interconnects.