Ignatenko E. Operation efficiency improvement for telecommunication nodes with cluster architecture

Object of research: telecommunication networks with cluster architecture Research objective: Efficiency increase for highly utilized nodes with cluster architecture. Development of nodes operation model and algorithms of incoming flow control. Research methods: statistical analysis, probability theory, queuing theory, teletraffic theory, fractal analysis theory, PC modeling. Theoretical and practical results, and novelty: mathematical models of: cluster node, arrival and handling of info-requests in telecommunication network were developed. As opposed to existing methods this one is considering features of cluster handling and request type, what gives more precise evaluation of network parameters; short-term prognosys method is being enchanced, which applies to info-stream behaviour in clustered network environment. As a differentiation to existing ones this uses monitoring algorithm of nodes state, statical characteristics of stream, request classification, floating prognosys horizon. Adaptive algorithm of nodes monitoring was developed, which allows to consider fluctuation of incoming stream and decrease the amount of overhead information. Dynamic algorithm of incoming stream control in clustered networking environment was developed. It considers predicted incoming flow behavior which allows to increase efficiency in high utilization conditions Software was developed for modeling of dynamic incoming stream control in clustered network environment, performance measurement and mean response time evaluation.Software was developed for adaptive monitoring control and incoming stream prediction. Methods of time series analysis for http requests stream was developed which has self-similar characteristics. Degree of implementation: Media-content network "MTS-Ukraine", virtual data storage "Telenet", Data center "Soniko-svyaz". Sphere of use: technical science, telecommunications.