Exploiting Relative Consensus Techniques in Future Advanced Communication Networks in the Presence of Failures

IIn this paper, a relative consensus algorithm is proposed to enable distributed systems to confront malicious behavior of the network nodes. In service networks, nodes get services from each other, and after each service provision, the receiving node calculates an indicative value of the quality of the received service. However, every node can behave maliciously, in either the service provision or the value calculation step. Both Byzantine and Logical malicious behavior types are considered. The aim is to identify the malicious nodes, and then to punish them by other nodes. This is achieved by exploiting Exponential Information Gathering (EIG) data structures. A new algorithm is formulated based on the EIG in order to reach a relative consensus agreement about the faultiness of one specific node before punishing it. The proposed algorithm is theoretically evaluated and discussed to confirm its correctness and validity. Moreover, it is shown that the proposed algorithm puts no constraints on the network structure, and since it has low implementation complexity, it is efficiently applicable to all network types including wireless networks