The experience of massive data processing in the cloud using Windows Azure (as an example)

The methods of using cloud computing technology to develop software that implements parallel processing distributed data are discussed in this paper. The basic resources of cloud systems are described. The virtual software-hardware complex is describes that implements a parallelization on the query processing level in the cloud system. The method of parallel implementation of a JOIN operation on the basis of symmetric horizontal distribution is considered. It is shown how one can build a virtual software-hardware complex for this operation by means of Windows Azure. Experimental data are presented that confirm the effectiveness of the proposed approach.