Design And Implementation Of Traffic Hot Spot Analysis System In Data Center Network

The rapid development of cloud computing and mobile Internet,has taken the Internet into the big data age,and the data center has developed very quickly.Along with scale of the data center network expand,the network applications is increaseing continuesly,and the network traffic is becoming more and more complexed,and the hot spots at the key point of the network is easily occured,giving much affects to the performance of network application service.Do a good job of analysising and reresearching the hot spots of the traffic of the data center network is very important to finding the hot spots in the data center network,and it is important to avoid of hot spots appear again by taking appropriate measures.Therefore,it is of great theoretical and practical significance to strengthen the analysis and research of the hot spots of the network traffic.The main innovations of this paper are as follows:(1)In view of the large scale of the data center network and the actual situation of the traffic flow,the sFlow technology is selected to sample the data center network traffic.Through the analysis of the sampled data,according to the distribution of the network flow,put forward a based on sFlow network traffic adaptive sampling method ASMNT,to solve the fixed sampling rate which is not according to the problem of the flow distribution and adjusting the sampling rate.(2)In the data center network terminal traffic prone to hot issues,we propose an analysis method of NTHSH-IE network traffic hot host based on information entropy.By calculating,we obtain the network traffic feature distributions of the hosts,using information entropy method of distribution characteristics of the flow are calculated and get the distributions of the size of the flow in,according to the distributions of the flow characteristics,judge in the current network is the emergence of a host,for the analysis of hot spot traffic host causes and avoid terminal flow too aggregation gives important reference.(3)In view of the problem that the traffic load imbalance in the data center network leads to a hot link,a method of analyzing the network traffic hot spots based on principal component analysis which we called the NTHSL-PCA method.Obtained by calculating the flow of information between hosts in the network.According to the principal component analysis method to the flow of dimensionality reduction and remove the correlation between flow and the host between traffic and routing information network traffic has been in the link distribution.In the end,the results of the analysis of the hot host and judge on a link in the presence of a hot spot traffic,on the formation of load balancing and flow scheduling decisions and avoid single hot link,which would give advices to optimize the network traffic distribution.(4)Design and implementation of a data center network traffic hot spot analysisprototype system.First based on the formal work we proposed the prototype of the system demand analysis and the overall design further detailed design and realize the main function modules of the prototype system,and realize the goal of data center network hot spot traffic analysis and results visualization display,network flow distribution scheduling and optimization to provide an important basis.