Research On Virtual Cluster Resource Allocation Of Multi-tenant Data Center

With the development of various applications on the Internet,in the era of cloud computing,the size of the data has long been more than the range of a single device processing capacity.Because of large-scale data under the various applications,especially Internet-based Internet applications,physical equipment resources such as CPU resources,network bandwidth,data storage and other consumption are increasing,and the data processing scale with the development becomes increasingly large.Data center as the current mainstream data upload,computing and storage devices,can provide more than the stand-alone resource supply capacity but clusters as resource providers,providing excellent hardware support for a variety of Internet applications.For any data center,the available resources are limited.Therefore,under the premise of satisfying the cluster demand,it is the research hotspot in the field of data center at home and abroad to realize the optimal allocation strategy.The above research hotspots include network scheduling,routing path allocation,network topology,graph theory and virtualization.This paper presents a resource allocation strategy for multi-tenant cloud data centers.The main goal is to provide a virtual cluster environment for users to provide cloud services.The resource allocation strategy adopted by this method will lay the network topology of the virtual cluster more rationally in the data center through the reasonable resource mapping algorithm,realize the traffic isolation of the tenants,guarantee the network bandwidth of different tenants,and reduce the consumption of physical resources.This method can effectively reduce the resource consumption and reduce the service cost for the cloud service provider.The main work of this paper is as follows:(1)Design a cluster resource providing solution for virtual data center based on OpenStack and SDN technology and implement a VDC cluster management system.Customers can send resource application to the system.The resources include virtual data center network topology,resources Quota demand and so on.The logical control layer of the system collects and processes the data of the user and obtains the optimal topology placement strategy,and sends the whole topology to the resource supply layer in the form of instruction set.(2)Design and implement a set of traffic separation system,SDN technology and OpenFlow protocol into the resource management layer,the OpenStack network model transformation,the network bandwidth according to different tenants were isolated,so that different tenants of the network chain Road decoupling,to prevent the occupation of bandwidth between different tenants,to prevent the interaction between the tenants of each other,to protect the tenant's network QoS and prevent data center internal traffic attacks.(3)related algorithm research,put forward a more universal model and implement a virtual cluster resource mapping algorithm VC-SE.In this algorithm,to achieve the minimum consumption of the actual physical resources to ensure optimal link to protect the entire physical data center can be maximized to be used.In this paper,by proposing a virtual environment resource management system for cloud environment,the ultimate goal of system design and implementation is to shield the underlying resources for users and to build a complete set of cluster environment quickly and easily.The system integrates the resource mapping algorithm VC-SE and introduces the OpenFlow protocol for bandwidth isolation and QoS guarantee under the multi-tenant,so that the system can make more efficient use of resources such as bandwidth in the process of virtual cluster creation.Reduce the traffic load in the data center.The next phase of the study is to optimize the energy consumption of the data center and to continually improve the problem and further improve the functionality of the system.