Authors: Cong Wang; Fengli Zhang
Addresses: School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, Sichuan Province 611731, China ' School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, Sichuan Province 611731, China
Abstract: To enhance the survivability of a network coordinate system (NCS) in an unsafe environment, the physical meaning of the latency estimation error is re-explained as the reputation of reference node instead of spring force. Thus, a vote-based model for network latency estimation is proposed, and categorised as a kind of l1 loss function optimising method. By taking the non-differentiability of the loss function into consideration, we introduce a variant of Rosenbrock algorithm to minimise this function and prove its convergence. The experiments show that compared with the traditional model, our model can achieve not only higher accuracy in benign environment but also better robustness in serious unsafe environments with low computing cost, and shows a strong intrusion tolerance capability.
Keywords: intrusion tolerance; NCS; network coordinate systems; network measurement; optimisation; vote-based models; network latency estimation.
International Journal of Computer Applications in Technology, 2015 Vol.52 No.2/3, pp.135 - 141
Published online: 26 Sep 2015 *Full-text access for editors Access for subscribers Purchase this article Comment on this article