PROJECT TITLE :

Optimization of the Processing of Data Streams on Roughly Characterized Distributed Resources

ABSTRACT:

The AS4DR (Adaptive Scheduling for Distributed Resources) scheduling method presented in this paper aims at maximizing throughput, when processing several information streams by divisible load applications on star-formed distributed memory platforms, with available speeds for communicating and computing which may be poorly estimated, or varying over time. The whole workload is supposed to be unknown. Consistent with the computation cost model, AS4DR will either maximize throughput, or CPU utilization by avoiding data-starvation of the computing units. An experimental assessment of the variation of the workload distribution to the variation of the communicating and computing speeds has been performed that shows that the employment of AS4DR will considerably improve the throughput. This paper additionally experimentally assesses a resource choice methodology to line up star-shaped clusters of distributed resources, thus as to method efficiently a set of knowledge streams with AS4DR.