Short-Term Spatio-temporal Forecasts of Web Performance by Means of Turning Bands Method

This work presents Turning Bands simulation method (TB) as a geostatistical approach for making spatio-temporal forecasts of Web performance. The most significant advantage of this method is requirement for the minimum amount of input data to make accurate and detailed forecasts. For this paper, necessary data were obtained with the Multiagent Internet Measuring System (MWING); however, only those measurements of European servers that were collected by the MWING's agent in Gdansk were used. The aforementioned agent performed measurements (i.e. download times of the same given resource from the evaluated servers) three times every day, between 07.02.2009 and 28.02.2009, at 06:00 am, 12:00 pm and 06.00 pm. First, the preliminary and structural analyses of the measurement data were performed. Then short-term spatio-temporal forecasts of total downloading times for a four days ahead were made. And finally, thorough analysis of the obtained results was carried out and further research directions were proposed.

[1]  Paul Barford,et al.  A Machine Learning Approach to TCP Throughput Prediction , 2010, IEEE/ACM Trans. Netw..

[2]  Qi He,et al.  On the predictability of large transfer TCP throughput , 2007, Comput. Networks.

[3]  U. Gerdtham,et al.  Relationship between exports, imports, and economic growth in France: evidence from cointegration analysis and Granger causality with using geostatistical models , 2011 .

[4]  Tevfik Kosar,et al.  A Data Throughput Prediction and Optimization Service for Widely Distributed Many-Task Computing , 2011, IEEE Trans. Parallel Distributed Syst..

[5]  Leszek Borzemski,et al.  Architecture of Multiagent Internet Measurement System MWING Release 2 , 2009, KES-AMSTA.

[6]  Leszek Borzemski,et al.  Distributed Web Systems Performance Forecasting Using Turning Bands Method , 2013, IEEE Transactions on Industrial Informatics.

[7]  Leszek Borzemski,et al.  CLIENT-PERCEIVED WEB PERFORMANCE KNOWLEDGE DISCOVERY THROUGH TURNING BANDS METHOD , 2012, Cybern. Syst..

[8]  Hans Wackernagel,et al.  Multivariate Geostatistics: An Introduction with Applications , 1996 .

[9]  Yong Wang,et al.  Spatial distribution of floating car speed , 2012 .

[10]  Leszek Borzemski,et al.  Internet Path Behavior Prediction via Data Mining: Conceptual Framework and Case Study , 2007, J. Univers. Comput. Sci..

[11]  G. Matheron.  The intrinsic random functions and their applications , 1973, Advances in Applied Probability.

[12]  Christian Lantuéjoul,et al.  Geostatistical Simulation: Models and Algorithms , 2001 .

[13]  Leszek Borzemski,et al.  THE EXPERIMENTAL DESIGN FOR DATA MINING TO DISCOVER WEB PERFORMANCE ISSUES IN A WIDE AREA NETWORK , 2010, Cybern. Syst..

[14]  Roger P. Karrer.  TCP Prediction for Adaptive Applications , 2007 .

[15]  Leszek Borzemski,et al.  MWING: A Multiagent System for Web Site Measurements , 2007, KES-AMSTA.

[16]  Kien A. Hua,et al.  Semi-Split TCP: Maintaining End-to-End Semantics for Split TCP , 2007 .