Next: References Up: A distributed approach to Previous: Performance

Summary

A distributed algorithm has been developed for high-performance information retrieval from a subject-specific corpus of information objects. A prototype system has been developed to measure the performance characteristics of the algorithm, and the prototype has achieved a throughput of 500 queries per second. The algorithm uses a local memory model and connectionless communication both of which are very useful for scaling up to larger corpora. The algorithm supports several levels of service and can tolerate a small number of communication and processor failures.


kenb@ccs.neu.edu
Fri Jan 20 21:47:36 EST 1995