Experimental Study of the Cloud Architecture Selection for Effective Big Data Processing

Abstract

Big data dictate their requirements to the hardware and software. Simple migration to the cloud data processing, while solving the problem of increasing computational capabilities, however creates some issues: the need to ensure the safety, the need to control the quality during data transmission, the need to optimize requests. Computational cloud does not simply provide scalable resources but also requires network infrastructure, unknown routes and the number of user requests. In addition, during functioning situation can occur, in which you need to change the architecture of the application - part of the data needs to be placed in a private cloud, part in a public cloud, part stays on the client.