Simulation Results – Energy Efficiency Sample Clauses

Simulation Results – Energy Efficiency. | | In our simulations, we consider master key loading schemes where R keys are distributed randomly6 among n clients such that any client possesses any master key with probability β. Figure 4 compares the average number of public multicast transmissions required for group key agreement by Algorithm 1 when n = 50, R = n/β , and β varies. Observe that as β increases, the number of transmissions required for key agreement decreases. This is consistent with intuition: as the occupancy set sizes increase, the minimal connected subhy- pergraph solution size decreases. To highlight the sublinear growth achieved by our protocol, the linear growth exhibited by Xxxxxxxxx and Xxxxxxx’x protocol [12] is also shown. The difference in performance between the two protocols can be attributed largely to our use of session keys derived from preloaded master keys – i.e., the BD protocol instead generates all keys on-the-fly. The MCSH problem assuming a random master key dis- tribution is closely related to random instances of the Set Cover problem. Toward connecting the two, consider the Set Cover problem with ground set X = {1, 2, . . . , t} and subsets Si ⊆ X, i = 1, . . . , R. A (β, t, R)-random instance of the 6When required by Lemma 1 for a given group, pairwise keys are added to the random distribution. Master key distributions derived from combinatorial designs are an alternative approach to providing random-like key distributions that satisfy Proposition 1. Combinatorial designs have been studied extensively in the context of the sensor network key distribution problem (see, e.g., [35]). Singer difference sets in particular are well-suited to our protocol. | |