Enhancements of the IEEE 802.11 Medium Access Control ... - PATS

12.3. Performance evaluation of the PHY-MAC cross-layer protocolfairness, because (probably) some of the nodes that have captured the channel toolong before, now do not have enough energy (or no energy anymore) to occupyaggressively a channel (in T2).In the last period of time, T3, we wanted to investigate the performance of ourprotocol in the next 50 seconds when DCF nodes are already dead. Here, thebehavior of the SNAPdMac protocol is investigated up to 350 seconds. Interestingto see is, how long a network with the SNAPdMac protocol can be alive andhow much its performance can be improved comparing with the DCF protocol (atthe moment of its death). In period T3 the total number of packets received gainobtained by the SNAPdMac protocol is increased up to 80%. However, note thatthis number will increase yet, because many of the nodes still live at 350 secondsas it can be seen in Fig. 12.8. At this point, it is important to mark that a fist nodedies at 343.42 seconds only (which increases the FND performance by 66% overthe standard). The throughput fairness still remains more than 130% better thanthe standard, where the sending bitrate decreases a little (from 81% till 53% gainover DCF). It is also interesting to remark that the Receiving LND is better thanthe network lifetime which means that nodes can still receive packets but theirintended transmissions cannot be realized, because routes are (probably) brokenor there is not enough total energy to complete their transmissions.Gain over DCF at 200, 300 and 350sec.225%175%125%75%25%Agg FND (sec):DCF = 207.17SNAPdMac = 343.42Agg Rcvd LND (sec):DCF = 211.69SNAPdMac ≈ 350Agg LND (sec):DCF = 222.49SNAPdMac ≈ 350-25%PacketRCVDΚ FND Lifetime LifetimeRCVDSbitfairnessThr.fairnessDelayT1- SNAPdMac vsDCF at 200sec.T2- SNAPdMac at300sec.T3- SNAPdMac at350sec.Figure 12.9: General results of HD/C scenario (1) - homogeneous network239