MSS S i Half Reply Other MSSs ACCEPT OR REJECT MH H i NEWTKT REQUESTACK ACKs ART Start ART End AWT Start AWT End
Experiment4- throughput vs. height of the tree Environment: All the MSSs are arranged in a tree like structure and we change the height of tree consisting of MSSs. As we increase the height, the number of MSSs also increases. As we are using binary tree structure for simplicity, the number of MSSs will be 2n-1. The number of MHs is xed as 130 for these simulations.
Why? The throughput for tree-based protocol is much better as an inquiry for an E-ticket has to travel only to certain levels in the tree and for each new E-ticket an inquiry message from a MSS can propagate at most until root. The throughput in the two-phase algorithms low as compared to the tree-based protocol since the number of message exchange increase with the increase in the number of MSSs.
Experiment5 & 6- ART(AWT) vs. % of duplicate E-tickets Environment: we have intentionally generated duplicate E-tickets by varying the time of generation of the E-tickets from 15 ms to 30 ms. Our aim in this experiment is to vary the duplicate E-tickets submitted at different time and observe the performance parameters. We simulated TPEP for both phases separately to check the efficiency in terms of average response time of the algorithm in each phase and a comparison between the two-phase I, two- phase II and tree-based protocol is given.
Experiment5 & 6- ART(AWT) vs. % of duplicate E-tickets ART vs. % of duplicate E-tickets
Experiment5 & 6- ART(AWT) vs. % of duplicate E-tickets AWT vs. % of duplicate E-tickets
Experiment5 & 6- ART(AWT) vs. % of duplicate E-tickets Why? In TBP the request gets rejected when it reaches the root as the E-ticket was already appended earlier to the vTkts list at the root In TPEP the submission of a duplicate E-ticket is treated as a submission of a fresh E-ticket, and MSS sends an inquiry for the sequence number of the E-ticket to all other MSSs visited by the MH. It generates extra messages and the MSS has to wait for the responses from other MSSs which increases response time. We conclude from experiments 5 and 6 that the time for the validation of an E-ticket is spent more in communication among the MSSs as the number of duplicate E-tickets increases.
Conclusions This paper proposed algorithm guarantees validation of an E- ticket at least once and at most once. In addition to proposing algorithms for E-ticket validation,a strategy for restructuring a distributed algorithm for mobile computing environment is also demonstrated. Our tree-based algorithms provides better response time, throughput and waiting time than the at version of the algorithm. We have also taken hosts resource constraints into consideration.
Future Work Security Issue the issues where mobile hosts cheat. P2P domain in wireless ad hoc networking environments using mobile agents.
Comment Whether the number of E-tickets in the Environment will affect the outcome? The TBP System