The workload balancing ETL system basing on a learning machine
Marcin Gorawski, Rafał Wardas · 2010
Data warehouses users usually expects both: short response time and high level of data “freshness”. The LEMAT presented as the ETL process manager bases on a concept of a adaptive load balancing of queries and actualizations according to user changing needs. The LEMAT system uses new workload balancing algorithm that uses LMWB (Learning Machine-based Workload Balancing) with the advanced query classifier SVM (Support Vector Machine). Moreover the method of a LEMAT system adaptation is presented. This method bases on collection of changing work conditions characteristics and reactions to congestions.