METHOD, APPARATUS AND SYSTEM FOR OPTIMIZING DISTRIBUTED COMPUTATION FRAMEWORK PARAMETERS

摘要

The present invention applies to the field of IT technologies and provides a method, apparatus and system for optimizing distributed computation framework parameters. The method comprises: obtaining currently submitted parallel computation framework jobs; retrieving historical distributed computation framework jobs of the same type as the distributed computation framework jobs in a distributed computation framework historical running database, the distributed computation framework historical running database comprising execution information and configuration parameters of the historical distributed computation framework jobs; retrieving historical distributed computation framework jobs similar to the distributed computation framework jobs among the distributed computation framework jobs of the same type; and performing optimized configuration on the configuration parameters of the distributed computation framework jobs according to the configuration parameters of the similar historical distributed computation framework jobs. The method, apparatus and system enable rationalization, automation and self-learning of Mapreduce job parameter configuration.