Autonomic Management of User-Centric Cloud Services for Smart Cities

Cloud services have become a de facto tool to address the computing need in processing large scale data generated in smart city environment where the data is processed in a data processing pipeline (DPP) to extract real-time insights. Such environment typically generates data with fluctuating rates thus requires an adaptive resource management to achieve both cost and QoS optimization goals. We leverage the cloud resource and contract heterogeneity to minimize the cost of resource allocation and proposed a novel end-to-end autonomic resource management framework built around multiple models and mechanisms for the adaptive scaling of cloud resources for DPP services ensuring the cost optimal end-to-end QoS fulfillment under fluctuating workload.