Research On Network Management In Software-Defined Data Center Networking

There are several golden years of Internet developments,the innovations and changes among techniques step into everyone's daily life.Internet brings numerous technical innovations to commercial and manufacturing industry.New applications are emerging every day and Internet becomes a new basis like water and electricity.The development is based on computing and storage resources,for example,the data centers.As an important resource pool,a typical data center environment not only included server clusters,security devices and monitoring systems but also resource scheduling mechanism,communication protocols and reliability strategy.They work together to accomplish different computing and storage tasks.Nowadays,artificial intelligence and data mining are leading a new era of computer science.The tasks that distributed in data centers are changed with the technique focus changing.The real-time requirements and data volumes are increased and applications are far more complicated than before.Data centers need to be stable and high-throughput to meet such requirements and such goals are rely on a reliable communication networks.Without surprisingly,data center networks is an important infrastructure.The topology,protocol and network management approach are evolved with changing requirements to certify a highly effective,stable and secure communication way.When SDN technique is proposed,it gives data center networks another opportunity.The separated control plane and data plane gives the network a centralized control ability and global view.The programmable characteristic saved operation and maintenance staff from the complex configuration tasks.In a word,soft defined data center networks created a collaborative ecosystem oriented to a variety of needs.In this paper,we summary the evolution history from traditional data center networks to soft defined data center networks.Aiming at several issues,such as network congestion and TCAM resource bottleneck problem,we proposed solutions and verified by a number of experiments.Our contributions are mainly list below:First,we make comparisons and analysis of current works on data center network topologies and discuss the impacts that SDN brings to data center networks;Aiming at TCP incast,network delay and SDN southbound protocols,we summary several important works and their advantages and short comings;For the flow management solutions,we introduce several approaches based on path selection,rate control and task self-adaptive adjustment and discussed their benefits.Our work makes a complement of current surveys since there are littler efforts on such comprehensive analysis.And we hope that it can heple researchers works efficiently.Second,we proposed a Qo S management solution to resolve the network service quality problem.In our solution,we use SDN obtain the network performance index and proposed a novel path selection approach to bypass the potentially congestion path.The experiments verified our solution's effeteness.Third,since large-scale data centers are major infrastructures in the big data era.Therefore,a stable and optimized architecture is required for Data Center Networks(DCNs)to provide services to the applications.Many research use SDN(Software Defined Network)based Multipath TCP(MPTCP)implementation to utilize the entire DCN's performance and achieve good results.However,the deployment cost is high.In SDN-based MPTCP solutions,the flow allocation mechanism leads to a large number of forwarding rules which may lead to storage consumption.Considering the advantages and limitations of the SDN-based MPTCP solution,we aim to reduce the deployment cost due to the use of an extremely expensive storage resource-Ternary Content Addressable Memory(TCAM).We combine MPTCP and Segment Routing(SR)for traffic management to limit the storage requirements.And to the best of our knowledge,we are among the first to use the collaboration of MPTCP and SR in multi-rooted DCN topologies.To explain how MPTCP and SR work together,we use a four-layer DCN architecture for better description,which contains physical topology,SR over the topology,multiple path selection supplied by MPTCP,and traffic scheduling on the selected paths.Finally,we implement the proposed design in a simulated SDN-based DCN environment.The simulation results reveal the great benefits of such a collaborative approach.Fourth,for the joint optimization problem of network resource and task allocation,we proposed a hybrid strategy for load balacing in the control laryer.By resolving the controller placement problem,we use the multi controller to optimized allocation of control-ability.Meanwhile,we propose a partition soluton for trans-reginal data center networks.According to the partitions,we can obtain a reasonalbe size of task flow and alleviate the processing pressure in busy hours.