Information security in mobile ad hoc networks.

摘要

Mobile Ad hoc Network (MANET) is a non-centralized, self-configuring network of mobile nodes connected by wireless links, the union of which forms an arbitrary and ever-changing topology due to the mobility of the nodes and the availability of the wireless links. Such unique features pose a number of security challenges not only from a harsh communication environment due to the factors of node mobility, radio fading and interfering, but also from the adversary's possible strength of eavesdropping, impersonating and even corrupting legitimate nodes in the wireless environment, where centralized online security authorities are not available. Aiming to provide information assurance that matches MANET's characteristics, this dissertation mainly focuses on two sectors within the security spectrum. The first is to design and analyze a set of unconventional Message Authentication Codes including Approximate Message Authentication Codes (AMACs) that can tolerate incident noise, and Division Butterfly Message Authentication Code (DBMAC) that has the capacity of error localization. The second is to design and develop a mechanism to provide Survivable and Secure Distributed Services (SSDS) in MANETs.; Different from conventional "hard" message authentications that are designed to detect even the slightest changes in messages, binary and N-ary, (N > 2), AMACs are designed to tolerate a small amount of noise in messages for applications where slight noise may be acceptable, such as in multimedia communications through wireless channel. AMACs are distance-preserving codes; i.e., the distance between two authentication tags reflects the distance between two messages. However binary representation of N-ary alphabets may destroy the original distance information between N -ary messages. Thus N-ary AMACs are proposed to better deal with the messages with alphabet size N. DBMAC is another new construct of message authentication codes. Not only can it provide the message authentication functionality, but it also can localize a few errors in the message. (Abstract shortened by UMI.)