The Research Of Resource Management And Scheduling Technology In Iaas Cloud Computing Platform

Cloud computing is a commercial service-oriented computing model, which integrates different types of heterogeneous resources including physical and virtual resources into a virtual resource pool and provides users with various types of services on demand. As an application mode of cloud computing, IaaS directly offers services which are encapsulated by computing resources, storage space and network bandwidth to users. Resource management and scheduling technology is the core of cloud computing and plays a very important role in platform stability and service reliability.Cloud platform mainly provides two types of services to users, including computing-oriented services and storage-oriented services. For computing-oriented services, system must be able to schedule the requests of users to reasonable virtual machines in the purpose of rapidly responsing to users while improving resource utilization, which requires a powerful reaource scheduling technology as a support. For storage-oriented services, system must ensure a quick read of the data and improve the reliability of resources and system throughput, which requires effective storage and management within the reources in cloud.In this paper, we design a generic resource management and scheduling framework which is named as HMAS based on Agent and CDN technology to simplify the complex cloud computing environment. Based on this, taking the response time of requests from users, resource utilization rate and system load balance into consideration, a new resource scheduling algorithm DGS which is based on improved greedy strategy and dynamic feedback resource scheduling is proposed to address the weaknesses of the previous scheduling algorithms. In view of the lack of static copy strategy which is widely used in mainstream cloud storage systems, we put forward a multiple active copy mechanism based on dynamic copy strategy which is named as MACM. Then an experimental environment is set up for simulation and the experimental results show that the proposed algorithms and mechanisms have great effect on improving service quality and guaranteeing the stability of system platform.