eduzhai > Applied Sciences > Engineering >

Superiority of Simplicity A Lightweight Model for Network Device Workload Prediction

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: The rapid growth and distribution of IT systems increases their complexityand aggravates operation and maintenance. To sustain control over large sets ofhosts and the connecting networks, monitoring solutions are employed andconstantly enhanced. They collect diverse key performance indicators (KPIs)(e.g. CPU utilization, allocated memory, etc.) and provide detailed informationabout the system state. Storing such metrics over a period of time naturallyraises the motivation of predicting future KPI progress based on pastobservations. Although, a variety of time series forecasting methods exist,forecasting the progress of IT system KPIs is very hard. First, KPI types likeCPU utilization or allocated memory are very different and hard to be expressedby the same model. Second, system components are interconnected and constantlychanging due to soft- or firmware updates and hardware modernization. Thus afrequent model retraining or fine-tuning must be expected. Therefore, wepropose a lightweight solution for KPI series prediction based on historicobservations. It consists of a weighted heterogeneous ensemble method composedof two models - a neural network and a mean predictor. As ensemble method aweighted summation is used, whereby a heuristic is employed to set the weights.The modelling approach is evaluated on the available FedCSIS 2020 challengedataset and achieves an overall $R^2$ score of 0.10 on the preliminary 10 testdata and 0.15 on the complete test data. We publish our code on the followinggithub repository: this https URL

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×