Research On Virtual Resource Allocation With Latency Awareness In Cloud Computing

Cloud Computing achieves wide-ranging concern,which becomes more and more important with the new concept "computing as a resource".There are three types of resources in cloud datacenters:computing,network and storage.There are thousands of racks existing in the datacenters.It is crucial to schedule these cloud resources and ensure the QoS for the cloud users.An enough effec-tive scheduling scheme can save plenty of communication cost,monetary over-head and electricity for enterprises and individual users.Optimizing network by allocating resource in cloud is a hot topic in the industry and academy area,and it becomes more complicated in the multi-cloud environment.Thus,this paper takes datacenter resource and user task into consideration,and puts the data transmission latency as optimal objects.Our research consists of three parts:The MapReduce latency optimization in standalone cloud,cooperative group latency optimization in multi-cloud,and virtual resource allocation with latency awareness in multi-cloud.The research optimizes latency by resource allocation,which is shown as follows:1).For MapReduce latency optimization in standalone cloud,we propose a latency optimization scheme based on virtual machine placement,which can significantly shorten the total data transmission latency and maximum data transmission latency.Firstly,we classify the virtual machines according to their latency with data nodes.Then,we place the Map layer virtual machines and optimize the total data transmission ans maximum data transmission latency respectively by the proposed algorithms.Finally,we place the Reduce layer virtual machine according to the number of virtual machine and threshold con-straint.The proposed scheme can effectively reduce the time complexity.The experimental results show that the proposed scheme can reduce 26.3%total data transmission latency and 20.6%maximum data transmission latency com-pared to the previous.2).For cooperative group latency optimization in multi-cloud,we propose a virtual machine rental scheme based on triangle inequality violation.A weighted graph is constructed to illustrate the multi-cloud network firstly.Then,we define a path latency model and a rental cost model for the cooperative group.After this,Yen’s algorithm and mixed integer programming are adopted to optimize the total data transmission latency with rental cost constraint.To study the performance of the proposal method,we conduct extensive evalua-tions using multiple real data traces and compare it with related approaches.The comprehensive evaluation analysis shows that the proposal achieves much lower total latency and is more flexible than other approaches.3).For resource allocation with latency awareness in multi-cloud,we pro-pose a novel resource allocation approach which aims to minimize the latency constrained by monetary overhead in the multi-cloud.We first define latency model and monetary overhead model.Then,we use K-means clustering to di-vide 22 datacenters into K regions.Finally,the region total latency optimization problem is modeled as binary quadratic programming,and we minimize the total latency of each region by placing the virtual machines.The evaluation is conducted with real data traces from multiple cloud providers.The results show that the proposed method can reduce 46.1%total data transmission latency ef-fectively compared with other approaches.