CiC (RDDC): Continuous Bulk Processing in the Cloud

点击翻译按钮获取中文摘要

This research explores an alternative data processing architecture thatfundamentally improves computing efficiency to reduce costs and provideenhanced data mining capabilities for cloud computing. The next majoradvancements in IT, medical, and science informatics will largely bedictated by our ability to store, manage, and analyze large amounts ofinformation. For many important problems, advances in data acquisitiontools are rapidly increasing data generation rates, exceeding our abilityto manage and process the data they produce.This proposal investigates novel architectures for continuous bulk dataprocessing, which support data-intensive applications that perform complexmulti-step computations over successive batches of input data (e.g., fromscientific or medical instruments), allowing analytics to simply beupdated, not recomputed, when new data arrives. The research seeks toraise incremental processing as a first-class abstraction, significantlylowering the barrier of data-intensive projects to take advantage of cloudcomputing. Towards that end, the work will develop data analysis portalarchitectures to increase the reach of on-demand data analytics acrossvarious application domains. As a case study, it will develop a portal forfatty-liver disease, allowing care givers on-demand access to medicalanalytics to improve care, reduce cost, and improve clinical outcomes forthis important disease.