Estimation of Level of Algorithms Parallelization and Efficiency of Using Different Error-Correction Codes in Cloud Realizations of Parallel Computing Algorithms

Abstract

The basic matrix algorithm for statement of parallel task in the cloud computing systems are considered in this paper. This algorithm is generally based on using the approach of arithmetic-logic relations and forming of recurrent matrixes. Possibility of using different error-coding methods for sending the request on parallelization task and receiving the answer is also considered. Comparative analyze of corrective ability of RS and convolutional codes is provided. As a result of provided investigations the relation for estimation the level of task parallelization is proposed. Choosing of convolutional codes for transmitting the data for parallelization the calculation task in cloud systems using of convolutional codes is recommended.