Performance characteristics of location-based group membership and data consistency algorithms in mobile ad hoc networks

Purpose - Many applications of mobile ad hoc networks require real-time data consistency among moving nodes within a geographical area of interest to function correctly, such as those that support disaster recovery and battlefield command and control. While it is operationally desirable to maintain data consistency among nodes within a large geographical area, the time and network resources required to propagate state changes to all nodes place practical limits on network size. This paper aims to investigate this issue. Design/methodology/approach - This paper investigates the notion of location-based data consistency in mobile ad hoc networks, and analyzes the tradeoff between data consistency and timeliness of data exchange among nodes within a location-based group in a geographical area of interest. Findings - Using a Petri net performance model, performance characteristics of location-based data consistency maintenance algorithms are analyzed and design conditions under which the system can tradeoff consistency for timeliness (reflecting the response time to propagate a state change) while satisfying the imposed data consistency requirement, when given a set of parameters characterizing the application in the underlying mobile ad hoc network are identified. Originality/value - The system designer can deploy the optimal area size identified in the analysis for maximizing membership and data consistency based on the "I am alive" beaconing interval used at runtime while satisfying the response time requirement for propagating a state change.