Towards Automated Cost-Efficient Data Management for Federated Cloud Services

Cloud computing has transformed the accessibility and usage of information technology resources, by offering them as services via the Internet. Cloud service provisioning spans across infrastructure, platform and software levels. The management of the services at these levels is based on monitoring. The analysis of monitoring data provides insight into Cloud operations in order to make informed decisions. Due to the emergence of numerous heterogenous Cloud platforms with proprietary APIs, service and monitoring data are being formatted using diverse and mostly incompatible data interchange formats. This results to interoperability issues and makes the analysis of monitoring data from multi-Cloud service deployments difficult to handle. The existing research efforts on data interchange formats have been mainly focused on general performance analyses. Little or no effort has been channelled towards a combination of multiple data interchange formats based on data type to achieve efficient serialisation that can facilitate interoperability in federated Clouds, and also reduce the size of data and bandwidth utilisation cost. This paper addresses these issues by presenting automated framework that is capable of automatically selecting the most suitable data interchange formats for achieving an efficient formatting and serialisation outcome. The goal of the framework is to enable robust and transparent communication within and between multiple Cloud deployments. Based on three use case scenarios, we evaluate the proposed framework to demonstrate its efficacy in formatting and serialising data.