Automated cloud resource orchestration
Realizing Infrastructure-as-a-Service (IaaS) cloud requires a control platform for orchestrating the provisioning, configuration, management and decommissioning of a distributed set of diverse cloud resources (i.e., compute, storage, network) serving different clients. Cloud resource orchestration is challenging due to the rapid growth of data centers, the high failure rate of commodity hardware, the enforcement of service and engineering rules, the increasing sophistication of cloud services, and the requirement to fulfill provider operational objectives and customer service level agreements (SLAs). Towards addressing these challenges, this dissertation makes following contributions: (1) An automated resource orchestration platform that allows cloud operators to declaratively specify optimization goals and constraints given provider operational objectives and customer SLAs. Based on these specifications, orchestration commands are automatically generated to optimize resource configurations and allocations within the cloud; (2) A highly available transactional resource orchestration platform for building IaaS cloud infrastructures. Transactional orchestration procedures automatically guarantee atomicity, consistency, isolation and durability (ACID) properties for cloud operations. Transactional semantics provide a clean abstraction which enables cloud operators to focus on developing high level cloud services without worrying about the complexities of accessing and managing underlying volatile distributed resources. We present the design and implementation of our transactional automated cloud orchestration platform. Using realistic scenarios and workloads derived from production cloud services, we demonstrate that our platform is able to automatically orchestrate compute, storage, and network resources within and across geographically distributed data centers to meet operational objectives and SLAs.
Liu, Changbin, "Automated cloud resource orchestration" (2012). Dissertations available from ProQuest. AAI3542826.