Software Defined Networking (SDN) is an emerging paradigm in computer networking that allows a logically centralized software program to control the behavior of the entire network. The SDN controller is a critical piece in this structure, where it is considered the mastermind of SDN networks. Thus, its failure will cause the entire network to fail. In order to overcome the problem of the overloaded controller failure in SDN, this project aims at proposing a controller offloading solution based on a prediction module that anticipates the presence of harmful long-term load. To predict whether the load in the controller is short-term or long-term load, we used three different classification algorithms: Support Vector Machine, k-Nearest Neighbors and Naive Bayes. Our evaluation demonstrates that Support Vector Machine algorithm is applicable for detecting the type of the load with an accuracy of 97.93 % in a real-time scenario.
PI "Redesigning Internet Protocols for Emerging Internet of Things Applications". KAUST CRG-5 with $1,487,649. (under review)
- N. Bouacida, A. Alghadhban, S. Alalmaei, H. Mohammed, and B. Shihada, "Failure Mitigation in Software Defined Networking Employing Load Type Prediction", IEEE International Conference on Communications (ICC), pp. 1-7, 2017. [PDF]
- Adma v1.0, 2016. Data mining tool to predict the SDN controller failures.
By downloading any of our software packages, you acknowledge that these software packages are provided for the research purposes only and are not permitted for commercialization purposes. Also, you are aware of the fact that additional support is not offered, nor authors liable under any circumstances. If you happen to use any parts of our software packages, you acknowledge to provide a correct referencing providing the software package URL.