The Community for Technology Leaders
Green Image
This article discusses Batrun, a system for distributed processing. We have developed a prototype of the system to automate the execution of jobs in a cluster of workstations. The objective is to use a general-purpose cluster, where machines are owned by different groups of users, as one virtual computer for batch processing. In contrast to a dedicated cluster, the scheduling in Batrun must ensure that only idle workstations are used for distributed computing and that local users, when they are operating, have full control of their machines. Batrun has two unique features: ownership-based scheduling policy to ensure top priority for owners of machines, multi-cell distributed design to eliminate a single point failure and to support scalability. The implementation of the system, based on multi-threading and remote procedure call mechanism, is simpler and more flexible than its predecessors.
batch processing, cluster of workstations, distributed processing system
Fredy Tandiary, Suraj C. Kothari, Ashish Dixit, E. Walter Anderson, "Batrun: Utilizing Idle Workstations for Large-Scale Computing", IEEE Concurrency (out of print), vol. 4, no. , pp. 41-48, Summer 1996, doi:10.1109/88.494603
185 ms
(Ver 3.3 (11022016))