Also we designed two decision algorithms, one for classifying the food categories and the other to classify and allocate/deallocate cloud instances in real-time. The work in [3] observed 40% improvement when the system had 80 parallel requests, as compared to 135.93% improvement in our system for the same set of parallel requests. In [21], another medical healthcare application is proposed that makes use of parallel image processing for image datasets from brain tumor specimens and developed pipeline analysis for study of brain cancers on hybrid cluster.