Presentation 2013-09-12
A Prediction Model of TCP Throughput for Intermittent TCP Access
Hiroshi YOSHIDA, Kozo SATODA, Tutomu MURASE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Fluctuation of TCP throughput on the Internet and mobile networks, which is caused by cross traffic and radio interference, spoils the user-perceived quality of cloud services. If the throughput fluctuation is predictable with a high accuracy, the service providers can be proactive to deal with the problem of the quality deterioration. The authors have constructed a high-accuracy prediction model of TCP throughput that is a mixture of a stationary process model and a non-stationary one on the basis of the stationary characterization. However, the prediction model requires continuous records of a TCP transfer (evenly-spaced time series data) such as video streaming. Therefore, the prediction accuracy is reduced when only intermittent records (unevenly-spaced time series data) are obtained, e.g., web browsing. In this paper, we elevate the time series model to a stochastic differential equation model that can predict the stochastic diffusion for the next three minutes on the basis of intermittent and unequally-spaced records with an accuracy of above 80%.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) TCP Throughput / Prediction / Stochastic Diffusion / Stochastic Differential Equation
Paper # IN2013-59
Date of Issue

Conference Information
Committee IN
Conference Date 2013/9/5(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Information Networks (IN)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Prediction Model of TCP Throughput for Intermittent TCP Access
Sub Title (in English)
Keyword(1) TCP Throughput
Keyword(2) Prediction
Keyword(3) Stochastic Diffusion
Keyword(4) Stochastic Differential Equation
1st Author's Name Hiroshi YOSHIDA
1st Author's Affiliation Cloud System Research Laboratories, NEC Corporation:Graduate school of Information Science and Engineering, Tokyo Institute of Technology()
2nd Author's Name Kozo SATODA
2nd Author's Affiliation Cloud System Research Laboratories, NEC Corporation
3rd Author's Name Tutomu MURASE
3rd Author's Affiliation Graduate school of Information Science and Engineering, Tokyo Institute of Technology:Cloud System Research Laboratories, NEC Corporation
Date 2013-09-12
Paper # IN2013-59
Volume (vol) vol.113
Number (no) 206
Page pp.pp.-
#Pages 6
Date of Issue