Impact of utilizing forecasted network traffic for data transfers

Sharing of information leads to the need to transfer data between geographically distant locations. Identifying the most appropriate time period to execute the data transfer is essential to achieve the best data transfer throughput; e.g. one can forecast network traffic, identify future low network traffic activities between two entities, and plan the data transfer accordingly. This forecasting can be done using Autoregressive Moving Average (ARMA) time series model. In this paper, we conduct an empirical study in a controlled laboratory environment to realise the impact of performing data transfer during the forecasted low network traffic activities. The network information is captured using a network analyzer and post-processed to create stationary data. This data is then passed to ARMA and the forecasting results produced by ARMA is post-processed to derive the forecasted network traffic activities. Comparison is made between the throughputs of the data transfers initiated when the forecasted network traffic is low and when the forecasted network traffic is high.