Faulty node repair and dynamically spawned black hole search in clouds

Abstract

The reality of the constant emergence of new threats justifies the necessity to protect network assets and mitigate the risks associated with attacks. In this context, eliminating faulty network entities in the distributed environment such as cloud of the clouds and smart grids catches the attention of the researchers. Among all threats, black hole is a severe and pervasive one which models a network site that disposes any incoming data without leaving any trace of such distraction. Black hole search is the process that leverages mobile agents to locate black holes in a fully distributed way. In this paper, we first review the state-of-the-art research in this area by categorizing the research results based on the adopted network models, being either synchronous or asynchronous. Most of the existing works focus on locating a single black hole. As for multiple black holes, the problem becomes even more complex. For the study of multiple black hole search, we introduce a new attack model that involves not only multiple faulty nodes in the network (a type of black hole), but also a gray virus that can again infect a previously repaired faulty node. Under such a model, the multiple faulty node search problem becomes more complex and realistic. We analyze the proposal model and identify key observations about the multiple faulty node search/location problem. We introduce one-stop and multi-stop gray virus and study the faulty node repair and black hole search problem. We first propose solutions that use a token model to solve the problem caused by a multi-stop gray virus in an asynchronous arbitrary network topology. Also under the token model, we then present solutions for the problem caused by this one-stop gray virus in an asynchronous ring network. Apart from the token model, we continue to study the problem caused by the one-stop virus using a whiteboard model, more particularly, with only one whiteboard in the homebase node in an asynchronous ring network. After proposing the new model and our algorithms, we conclude some future work on both single and multiple black holes search. We also highlight some open problems on the one-stop and multi-stop gray virus.