Rhizome: A middle-ware for cloud vision computing framework

Over the past decade, high performance applications have embraced parallel programming and computing models, like video analysis for surveillance purpose. However due to the heterogeneous, opaque and dynamic nature of Cloud Computing, existing Cloud parallel framework like Hadoop/HPCC (with Job-keeper + Map-reduce behind) approach is not suitable to perform vision computing specific task on the cloud. Rhizome is a middle-ware as PAAS & IASS for building specific cloud vision computing applications, so as to convert a high performance application to an elastic cloud one, which may enable effective high performance and robust parallel computing especially for streaming like task e.g. vision computing within the cloud environment. Rhizome encourages use of simple MoC based concurrency patterns that are easy to understand and implement correctly, and built on an agency-like model with declarative specification of persistence, replication, and consistency and using lightweight transactions in cloud OS environment to support the development of reliable and scalable client + cloud software. The experimental assessment demonstrates a superior performance of the cloud vision computing application in Hadoop with the help of Rhizome, which provides substantial and consistent better programming easiness and system robustness.