Modeling Server Workloads for Campus Email Traffic Using Recurrent Neural Networks
As email workloads keep rising, email servers need to handle this explosive growth while offering good quality of service to users. In this work, we focus on modeling the workload of the email servers of four universities (2 from Greece, 1 from the UK, 1 from Australia). We model all types of email traffic, including user and system emails, as well as spam. We initially tested some of the most popular distributions for workload characterization and used statistical tests to evaluate our findings. The significant differences in the prediction accuracy results for the four datasets led us to investigate the use of a Recurrent Neural Network (RNN) as time series modeling to model the server workload, which is a first for such a problem. Our results show that the use of RNN modeling leads in most cases to high modeling accuracy for all four campus email traffic datasets.
KeywordsEmail traffic Model server workload Recurrent Neural Network Time series modeling
We would like to sincerely thank Mr. Panagiotis Kontogiannis, Head of the Educational Computational Infrastructure at the Technical University of Crete, Mr. Martin Connell, Senior Systems Engineer at LJMU and Mr. Mario Pinelli, Manager of Computer Services and IT at Murdoch University. Without their help with collecting the datasets this research would not have been possible.
- 2.Takemura, T., Ebara, H.: Spam mail reduces economic effects. In: Proceedings of the 2nd IEEE International Conference on the Digital Society (2008)Google Scholar
- 3.Kashyap, A., et al.: Internet Security Threat report (2014). http://www.symantec.com/content/en/us/enterprise/other_resources/b-istr_main_report_v19_21291018.en-us.pdf. Accessed 15 June 2017
- 5.Bertolotti, L., Calzarossa, M.C.: Workload characterization of email servers. In: Proceedings of SPECTS (2000)Google Scholar
- 13.Boukoros, S., Kalampogia, A., Koutsakis, P.: A new highly accurate workload model for campus email traffic. In: Proceedings of the International Conference on Computing, Networking and Communications (ICNC), pp. 1–7 (2016)Google Scholar
- 14.Navaroli, N., DuBois, C., Smyth, P.: Statistical models for exploring individual email communication behavior. In: Proceedings of the Asian Conference on Machine Learning (2012)Google Scholar