A VDC Mapping Algorithm With Reliability-Related Network Features

In recent years,the rapid development of cloud computing and data center technologies,and the increasingly widespread application of data centers have made the existing traditional data center network structure unable to adapt to the rapidly expanding network scale.In order to use data center virtualization technology to improve the economic benefits of data center,cloud service providers abstract tenants'requests into Virtual Data Center?VDC?,and design a VDC mapping algorithm to map VDC requests to physical components,providing services for these VDC requests at the same time,so that multiple tenants can share resources.However,in the process of data centers virtualization,there will be a number of tenants'requests that are highly demanding on service reliability.Providers of cloud services must ensure the provision of stable services to these requests in order to reduce the loss of service failure caused by the failure of these requests.In addition to the fault-tolerant nature of the data center itself,it is also necessary to design a VDC mapping algorithm for the reliability of the mapping service.In the research of the VDC mapping algorithm in the past,there are also some studies on reliable VDC mapping.Most of these studies can perform reliable VDC mapping in one or several types of typical data centers without considering the features of the data center structure itself.Considering the structure of the data center itself,it can bring more ideas to optimize the reliability of VDC mapping.And now with the increasing number of data center types,the scale increasingly,the application of VDC mapping algorithm on different types of data centers is becoming more and more inconvenient.The above problems can be solved by designing a VDC mapping algorithm which can analyze the structure features of the data center and apply them.In this paper,we study the topology structure of some data centers and extracts relevant information.Using these structural features information,it analyzes which features have more influence on reliable VDC mapping,and it analyzes which feature servers and switching devices and links need to be used in the VDC mapping process,and prepares to integrate these reliability-related features into the VDC mapping algorithm to improve the reliability of the mapping results.We introduce a reliable VDC request model with backup nodes and gateway nodes called RVDC model at first.RVDC is equipped with backup nodes for providing the maximal reliable service and gateway nodes as the entrance of the VDC.Secondly,Measuring the reliability of a VDC mapping solution accurately has been proved as a NP problem,therefore we propose a method to estimate it.And then,we analyze the two influence factors which affect the reliable mapping,and quantified them into two network features,i.e.the distance between two nodes dis(₄),₅))and the number of the shortest path between two nodes pn(₄),₅)).At last,we propose a VDC Mapping Algorithm with Reliability-Related Network Features?abbr.RRNF-VDC algorithm?,which is a reliable VDC mapping algorithm based on data center reliability-related network feature analysis.The RRNF-VDC algorithm combines two reliability-related network features and summarizes the optimization operators which can affect the reliability of mapping.According to the optimization operators,the candidate nodes in the data center are sorted to select the set of nodes with higher reliability of mapping.The RRNF-VDC algorithm use RVDC model,considers that the impact of the backup server and the backup switching link on the reliability of the mapping service.The mapping process is divided into three main steps:mapping of virtual gateway nodes,mapping of other virtual nodes and mapping of virtual links.At the same time,this paper uses the control experiment to verify that the RRNF-VDC algorithm does improve the reliability of the mapping and for the general data center structure,the results of the reliability-related network feature analysis are all valid.