A high-efficiency data center cloud server resource autonomous management method

Abstract

The invention provides a high-efficiency data center server resource autonomous control method, which belongs to the field of computer high-performance computing. The system includes a resource real-time scheduling layer, an energy consumption control layer and a virtual machine migration layer. The resource real-time scheduling layer includes a performance monitor, an online estimator A, a resource controller, and a resource allocator. The energy consumption control layer includes a performance monitor, an online The estimator B, the energy consumption controller, the frequency regulator, and the virtual machine migration layer include a performance monitor, a load detector, and a virtual machine migration scheduler. The method includes three parts: resource real-time scheduling, energy consumption control and virtual machine migration, which are respectively coordinated at the level of seconds, minutes and hours, so that the data center server can achieve optimal control of performance and energy consumption. The method and system can realize the full-cycle multi-level cloud resource scheduling management of the data center server, greatly reduce the energy consumption of the data center server, and ensure the quality of cloud services.

Description

technical field [0001] The invention relates to a method for self-management of cloud server resources in a high-efficiency data center, belonging to the field of computer high-performance computing. Background technique [0002] With the further prosperity of the Internet economy, the scale of the data center is expanding at an alarming rate, especially with the introduction of cloud computing and big data, the scale of the data center has been developed unprecedentedly, how to use the resources of the data center (especially cloud servers) Efficient management is a challenging problem. At the same time, while data centers consume a lot of energy, they also bring carbon emissions that cannot be ignored. Therefore, it is of great significance to design an efficient and reasonable cloud server resource control method and system for the resource management of the entire data center, as well as to reduce the energy consumption of the data center, reduce operating costs, and re...

AI Technical Summary

Problems solved by technology

[0004] With the rapid development of cloud computing technology, relevant researchers have proposed various cloud server performance management solutions. Although these solutions can optimize cloud server performance and reduce cluster energy consumption to a certain extent, they are often controlled by a single control component or Scheduling algorithm composition, the designed system lacks one-stop full lifecycle management

At the same time, most of the existing control schemes are deterministic control, relying on the assumption that the online estimator can stably provide accurate model parameters to the controller

However, this assumption is often not true, because due to the complexity of the data center cloud server system, the linearization method is usually used for modeling, and the nonlinear factors in the real data center cloud server are often ignored in the modeling process (such as computer system Resources are limited and limited by the manufacturing process of the hardware itself; sudden web loads can cause unknown fluctuations in the output of the system; the randomness of the internal processing of the computer, etc.)

Therefore, when sudden web loads cause severe perturbations at the output of the system, the performance of the online estimator becomes inaccurate, leading the controller to make wrong control choices), and in some cases, even causing the application's response Oscillations in time

Images