Orchestration to Optimize Energy and Demand
摘要
Following an exploration of opportunities and flexibilities of demand-side management (DSM) within data centers in the preceding chapter, this chapter focuses on demonstrating the effective utilization and implementation of DSM through software-based orchestration. Section 5.1 details the essential process of implementing the feedback control loop, providing a comprehensive explanation of each step involved. In Sect. 5.2, the energy management system (EMS) is introduced as a software-based orchestrator, outlining its key tasks and corresponding components. Furthermore, a multi-agent approach is proposed, integrating the contract net protocol (CNP) to enhance coordination among the agents within an EMS. The latter sections of this chapter highlight two primary optimization objectives: reducing energy consumption and adapting power demand. Section 5.3 introduces an EMS with a primary focus on decreasing energy usage while upholding the service level agreements (SLAs) of IT customers. It commences with an overview of the foundational elements of the EMS architecture, followed by an in-depth exploration of its components. Additionally, a template is introduced to effectively describe the ICT resources of data centers, accompanied by a practical example for clarity. Section 5.4 addresses the aspect of demand adaptation within EMS, introducing concepts such as relaxed SLAs and flexible supply-demand agreements (SDAs), along with straightforward reward and penalty mechanisms, supported by a real-world illustration. The chapter concludes with a summary in Sect. 5.5.