Cost Implications for Survivability of Terrestrial Networks Under Malicious Failure

Protection against malicious failure is a very important issue in many network applications. While no general theory is easily developed, the following general statement emerges from any analysis: the combined link cost (topology) to guard against such failure is much higher than to protect against random failure. In this paper we first define precisely the concept of malicious failure and immediately show the severe cost implications. To constrast this a summary of the major results in random failure is provided. We then manually carry out some survivability computations on networks on a small number of nodes; we also address the issue of partial knowledge.