Multi goal self-adaptation for micro-service platforms

点击翻译按钮获取中文摘要

Cloud computing has become a software platform of choice for both consumers and business. Deployed as public, private or hybrid, a cloud provides flexibility, cost control and scalability. Hybrid clouds try to exploit and balance between the security, privacy and compliance offered by private clouds, and the economies of scale provided by public clouds. This project focuses on applications deployed on container-based hybrid clouds and investigates application deployment and management patterns, as well as efficient and privacy-preserving elasticity of those environments. The type of applications we consider are those that rely on big data technologies that dominate areas such as Analytics, Internet of Things (IOT), or cyber physical systems (CPS). The project has the following long-term goals: The first goal of the project is to research and build efficient models for micro-services on hybrid clouds. Micro-services are small units of program that reside in containers, which in turn are units of deployments in virtual machines. The second goal is to investigate and evaluate runtime deployment and optimization algorithms for efficient application elasticity in container-based hybrid clouds. We are going to focus on algorithms that address cost, performance and privacy and which optimize micro-service deployments in hybrid clouds. The third goal is to investigate, implement and evaluate techniques for data partitioning among private and public clouds. We consider NoSQL datastores and analytics workloads.