13.07.2015 Views

Page 2 Lecture Notes in Computer Science 2865 Edited by G. Goos ...

Page 2 Lecture Notes in Computer Science 2865 Edited by G. Goos ...

Page 2 Lecture Notes in Computer Science 2865 Edited by G. Goos ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Analyz<strong>in</strong>g Split Channel Medium Access Control Schemes 13710.90.8Throughput of MAC−1 and MAC−2R0.70.60.50.40.30.20.1L d=1024, S 1L d=1024, S 2RL d=1024, S 2R, simulationL d=2048, S 1L d=2048, S 2RL d=2048, S 2R, simulationL d=4096, S 1L d=4096, S 2RL d=4096, S 2R, simulation00 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5Ratio of control channel over entire channel, rFig. 5. Throughput comparisons between MAC-1 and MAC-2RIn Fig. 5, we compare the throughput performance of pure ALOHA-basedMAC-1 and MAC-2R schemes for different data packet lengths. The straight l<strong>in</strong>esrepresent the throughput of the MAC-1 scheme. The throughput of the MAC-2R scheme <strong>in</strong>creases as r <strong>in</strong>creases until the throughput reaches the maximumachievable value and then degrades. When r is too small, the control subchannelneeds much longer time to come up with a successful RTS/CTS dialogue.However, when r is too large, the fraction of the entire available channel used totransmit data is too small, limit<strong>in</strong>g the throughput of the MAC-2R scheme.Compar<strong>in</strong>g the throughput performance of the MAC-1 and the MAC-2Rschemes, we observe that the MAC-1 scheme always out-performs the MAC-2Rscheme, due to the non-zero wait<strong>in</strong>g time on the data subchannel <strong>in</strong> the MAC-2R scheme. As expected, the throughput <strong>in</strong>creases as L d (or k) becomes larger,approach<strong>in</strong>g 1 as L d (or k) <strong>in</strong>creases. In the same figure, Fig. 5, we also draw thesimulation results of the MAC-2R scheme, demonstrat<strong>in</strong>g that our simulationresults closely match those obta<strong>in</strong>ed <strong>by</strong> our analysis.In Fig. 6, we show the ratio of the throughputs of the MAC-2R and the MAC-1 scheme, S 2R /S 1 , as a function of r for different data packet lengths L d .Itcanbe observed that the maximum achievable throughput of the MAC-2R schemeis closer to the throughput of the correspond<strong>in</strong>g MAC-1 scheme as L d <strong>in</strong>creases.Thus, the penalty for splitt<strong>in</strong>g the s<strong>in</strong>gle channel is lower when data packetlength is larger. As L d <strong>in</strong>creases, the optimum r that achieves the maximumthroughput for the MAC-2R scheme becomes smaller.In Fig. 6, we also draw symbols represent<strong>in</strong>g the performance of the MAC-2R scheme, when the s<strong>in</strong>gle channel is split accord<strong>in</strong>g to the expected value ofthe contention resolution periods. In these cases, r is set to r ∗ = w+2w+2+k ,asshown <strong>in</strong> (8). As shown <strong>in</strong> the figure, the throughput of the MAC-2R schemes isoffset from the optimum operation po<strong>in</strong>t of the MAC-2R scheme. Interest<strong>in</strong>gly,we f<strong>in</strong>d that such a non-optimum scheme would operate at the same relativeperformance S 2R /S 1 for the different values of L d , as the three symbols are all

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!