Energy-efficient Traffic Management Strategy In Software Defined Data Center Networks

The explosive expansion of data center sizes aggravates the energy consumption,which has restricted the sustainable growth of cloud services and seriously troubled data center providers.In recent years,plenty of advanced data center network architectures have been proposed,which adopt richly-connected topologies and multi-path routing technique to provide high network capacity.However,these frameworks have the low energy efficiencies in the network during the traffic valley time.To address the problem,the energy-aware flow routing / scheduling algorithm is proposed,the basic idea of which is to aggregate network flows into a subset of links / switches and put the idle links / switches into sleep mode for energy saving.In general,link capacity is the main constraint in the energy-aware algorithm.Redundancy Elimination(RE)is proposed in recent years which is a applied to increase the network link capacity.RE can remove duplicate content from network transmission,that is,whenever the router at the sending side notices the same data pattern in the data repository,it sends a small signature instead of the original data,and then the receiving router looks up the signature to recover the original data.However,it neglects the energy consumption on routers caused by RE.Therefore,this thesis designs a flow preemption routing scheme based on energyaware and redundancy elimination techniques at first.Then,a traffic consolidation algorithm is proposed to avoid network oscillation and network performance degradation when the modes of network devices are switched frequently.The main contents of this thesis are as follows:(1)We propose a flow preemption routing scheme with redundancy elimination(REFPR).The RE-FPR scheme adopts software defined networking(SDN)technology to select different routing paths for the traffic flow and control RE function on the corresponding router under two modes,i.e.,traffic peak and traffic valley.Specifically,the RE-FPR scheme also employs the SDN controller to update flow states and link states of the network.This thesis formulates RE-FPR scheme as a energy consumption minimization problem,which subjects to flow conservation constraint and link capacity constraint.Furthermore,we transform the optimization problem into a zero-one linear programming problem.Simulation results show that RE-FPR scheme outperforms the traditional flow routing algorithms in term of flow completion time and the number of active links / RErouters.(2)During traffic peak time,network device patterns are switched frequently,which can cause network oscillation and performance degradation.In order to solve this problem,we propose a correlation-aware traffic consolidation algorithm(called CATC)in energy-efficient data center network.In this algorithm,it processes and forwards network flows as far as possible,which can reduce the switching times among different modes and prolong the sleep time of network devices.Finally,we evaluate CATC algorithm in Mininet which is a network simulator.And the simulation compare CATC algorithm with general centralized solutions,and CATC algorithm not only saves energy upto 45% at most for the data center network,but also increases little network delay.