A cloud architecture approximation to collaborative environments for image analysis applications

Abstract

This work describes a flexible cloud computing architecture intended to be used in collaborative groups where each group member develops image analysis processes that are made available to the rest of the group in a flexible, robust and easy-to-use way. The cloud computing approximation makes the whole system elastically scalable and reliable to failures. Computing resources are provisioned when needed, used for a specific task and finally relinquished after the job is done. In the proposed architecture each individual process takes its input data from a queue, carries out the task and leaves the output in another queue. Building a complex task starting from individual ones is performed by chaining processes, just matching output and input queues of subsequent processes.