HTMTAD: A Model to Detect Anomalies of CDN Traffic Based on Improved HTM Network
There will always be malicious intrusion, node downtime and other events caused by network traffic anomalies while Content Delivery Network (CDN) is facing user’s service. These events will lead in a large area of network paralysis and suspension of network services. Therefore, in order to effectively detect and deal with the anomalies in advance, the paper makes a partial improvement on the existing Hierarchical Temporal Memory network (HTM), and proposes a new network model HTMTAD (Hierarchical Temporal Memory – based Traffic Anomalies Detection) to detect intelligently the changes of abnormal traffic from the CDN. In view of the characteristics of CDN traffic data, the paper proposes a hash coding algorithm to improve the reliability of encoder and an anomaly likelihood calculation method to detect the CDN traffic anomalies. Experimental results show that HTMTAD can effectively detect anomalies in CDN network traffic.
KeywordsCDN Traffic anomaly detection Hierarchical Temporal Memory Encoder Anomaly likelihood
The authors would like to thank the anonymous reviewers for their valuable comments and suggestions. This work is supported in part by the National Natural Science Foundation of China under Grant 61170035, 61272420 and 81674099, Six talent peaks project in Jiangsu Province (Grant No. 2014 WLW-004), the Fundamental Research Funds for the Central Universities (Grant No. 30916011328, 30918015103), Nanjing Science and Technology Development Plan Project (Grant No. 201805036), the Open Fund Project for Improve government governance capacity Big Data Applied Technology National Engineering Laboratory 2017–2018.
- 1.Etsy Skyline GitHub: Online Code Repos. https://github.com/etsy/skyline
- 3.Basseville, M., Nikiforov, I.V.: Detection of abrupt changes. Change 2, 729–730 (1993)Google Scholar
- 4.Angelov, P.: Anomaly detection based on eccentricity analysis. In: IEEE Symposium on Evolving and Autonomous Learning Systems, EALS, pp. 1–8 (2014)Google Scholar
- 5.Costa, B.S.J., Bezerra, C.G., Guedes, L.A., Angelov, P.P.: Online fault detection based on typicality and eccentricity data analytics. In: International Joint Conference on Neural Networks, pp. 1–6. IEEE (2015)Google Scholar
- 7.Netflix Surus GitHub: Online Code Repos. https://github.com/Netflix/Surus
- 8.Laptev, N., Amizadeh, S., Flint, I.: Generic and scalable framework for automated time-series anomaly detection. In: The 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 1939–1947. ACM, August 2015Google Scholar
- 9.Hawkins, J., Ahmad, S.: Why neurons have thousands of synapses, a theory of sequence memory in neocortex. Frontiers Neural Circ. 10, 23 (2016)Google Scholar
- 10.Ahmad, S., Hawkins, J.: Properties of Sparse Distributed Representations and Their Application to Hierarchical Temporal Memory (2015). arXiv preprint: arXiv:1503.07469
- 14.Hawkins, J.: Biological and Machine Intelligence. release 0.4 (2016). http://numenta.com/biological-and-machine-intelligence
- 15.Purdy, S.: Encoding Data for HTM Systems (2016). arXiv preprint: arXiv:1602.05925