Component-Based Software (September 1995)

Cost-Based Generation of Scalable, Reliable, Real-Time Software Components

Sponsor: Clearsight (formerly HyBrithms Corporation; Sagent Corporation)

• Project Performance Period: 9/1/1995 - 12/31/1997
• Total project (est.): $2,106,800.00
• Requested ATP funds: $1,938,800.00

Business and industry, both large and small, increasingly are relying on complex, distributed networks of computers and information systems. Manufacturing firms are studying "virtual enterprises" which could tightly connect, within computer networks, their design and manufacturing operations with their suppliers. International banking firms rely on networks to coordinate activities among far-flung branch offices in the volatile world of international finance. Independent physicians in a medical network, utilities, food and commodity distributors, retailers, and communications companies are all examples of industries that now or potentially can use complex computer networks to manage operations. Efficient and effective control of distributed systems is difficult, particularly in the increasingly common situation where the system is a "hybrid" including discrete and continuous evolution elements and digital and analog signals. A major obstacle to effective and efficient control of these systems has been the need for network and application synchronization--the task of assuring that events between different nodes on the network happen in the proper order and that data managed by different nodes remain consistent. Developing such synchronization has been a particularly intractable problem, but a recent breakthrough in theoretical mathematics offers the possibility of obtaining very fast solutions to large problems of this type. Sagent Corporation, whose principals were responsible for this advance, proposes to develop the basic algorithms of this theory. These will make possible scalable, reliable, real-time software components that can be used to build network "agents" that interact to synchronize and coordinate distributed systems, both large and small. If successful, the project will simplify the development and maintenance of complex real-time distributed systems, lowering the costs both of establishing and operating such systems. This technology could help overcome the current failure rate, estimated at between 70 and 90 percent, in building such systems.

For project information:

Dr. Wolf Kohn, (206) 637-1180
wk@sagent.com

ATP Project Manager

Barbara Cuthill, (301) 975-3273
barbara.cuthill@nist.gov