Multi-objective virtual machine adaptive position selection method and distributed cloud system

Abstract

The invention belongs to the technical field of virtual machines in cloud and discloses a multi-objective virtual machine adaptive position selection method and a distributed cloud system. By adoptinga fuzzy logic control parameter and using a fitness function of multi-dimensional cooperation, a search space of solving is optimized; based on a biological immunological memory mechanism, ordered coding is performed for populations according to groups and resources in a position selection process, and population updating is performed in combination with an epsilon dominant elitist strategy; anda tournament selection method is used during population selection, and the populations are subjected to optimization processing through a single-point operator crossover mechanism and an X point variation operation. For verifying the performance of an algorithm, PHM, NSG-2, DRF and MOVMLSA algorithms are subjected to simulation experiment; and an experiment result shows that the immune operator-based multi-objective virtual machine position selection algorithm has great advantages in aspects of resource utilization rate, physical machine load balance degree, unit resource cost and the like.

Description

technical field [0001] The invention belongs to the technical field of virtual machines in the cloud, and in particular relates to a multi-target virtual machine self-adaptive position selection method and a distributed cloud system. Background technique [0002] At present, with the increasing expansion of network application services, there is an urgent need to efficiently integrate information technology architecture and various resources, so as to effectively manage and control physical resources as a whole, improve resource utilization, and reduce unit resource utilization costs. Virtual machine technology is a key technology in the field of virtualization, and has been widely used in distributed clouds. The rapid popularization and wide application of cloud computing make people gradually approach the goal of unlimited reasonable application of limited resources. In this mode, physical resources can be like adding fuel and natural gas to cars in daily life, and users ...

AI Technical Summary

Problems solved by technology

However, if the goal is to reduce migration, it will inevitably increase the number of physical nodes used

For the virtual machine initialization problem, that is, the virtual machine static placement problem, the traditional genetic algorithm is used to deal with it. The disadvantage is that the overhead in the virtual machine migration process is not calculated.

In many cases, the deployment effect obtained is not the result that the user most expects

[0005] (3) Many new problems have emerged in the current dynamic allocation and management of virtual and physical shared resources in data centers

Since the influence of factors such as system performance, resource cost, and consumption must be comprehensively considered when selecting a virtual machine location, this will make the problem more complicated.

[0006] To sum up, the problem existing in the existing technology is: the existing virtual machine placement method in cloud computing does not select the most suitable data center from the distributed cloud; thus affecting the rational use of multi-dimensional resources

In most cases, only a certain dimension is optimized, so that the local optimal solution can not be obtained, and the global optimal solution cannot be obtained; there is a lack of an effective virtual machine adaptive location selection framework

Images