Resource Management Mechanisms for Mo- Mo-bile Broadband Networks
3.4 Experimental Results
3.4.2 Effectiveness of the Proposed Cross-Layer Resource Management SchemeScheme
SIP-MNG is disconnected from the Internet; the other is the online case, where SIP-MNG is connected to the Internet. In the offline case, for SIP-MNG, there is no SIP signaling cost when it moves from one network domain to another. The external push server would be SIP-MNG’s representative, which answers and transfers incoming sessions for the mobile network. On the other hand, since the MIPv6-NEMO scheme does not support the disconnected feature, the MR still has to track network signaling and update with its HA when handoff. Suppose that the HA binding update cost is costHABU. In the online case, the SIP signaling cost of SIP-MNG when handoff will be N × costSIP-reregistration+ S × costSIP-reIN V IT E, where N is the number of MSs in the mobile network, S is the number of sessions between the mobile network and the Internet. On the other hand, the MIPv6-NEMO scheme with routing optimization will require a cost of costHABU + M × costBU when handoff, here we assume that the routing optimization approach based on binding update for network prefixes is used, and M is the number of CNs that are communicating with the mobile network. To summarize, in the online case, the SIP has its disadvantage in terms of signaling cost as compared to the MIPv6-NEMO scheme with routing optimization. However, in the offline case, our approach incurs lower cost than the MIPv6-NEMO scheme, which may compensate some of the signaling cost in the on-line case.
3.4.2 Effectiveness of the Proposed Cross-Layer Resource Management Scheme
To verify the correctness and applicability of the proposed algorithm, an event-driven C++ simulator is developed. Unless otherwise stated, the following assumptions are made in our simulation. (1) The same call arrival model, call holding time, and call residence time as specified in Section 4 are used in the simulation. (2) Parameters of IEEE 802.11b and 802.11e are used. (3) We set TXOP limit to zero for four ACs, which means that a QSTA can only transmit one packet in each successful contention. (4) The communication channel is assumed to be error-free. (5) No RTS/CTS is used. (6) The BI=500 ms. (7) For AC VO traffics, we set CWmin to 7, CWmax to 15, and AIFSN[AC VO] to 2. For AC BE traffics, we set CWmin=31, CWmax=1023, and AIFSN[AC BE]=3. (8) Since
0
∗ : Blocking Rate (simulation)
: Dropping Rate (analysis)
◦ : Dropping Rate (simulation)
................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
Figure 3.25: Comparison of simulation and analytical results on blocking rate and drop-ping rate (λh = 0.8, λn= 1.2).
QAP is more likely to be the performance bottleneck, we give it a higher priority by setting its TXOP[AC VO] to N packets where N is the expected number of voice QSTAs in the QAP. (9) In QSTAs, each AC has a queue of size 50. (10) A voice packet will be dropped when it is buffered in a queue for more than 100 ms. (11) A packet can be re-transmitted at most 3 times. (12) G.726 with 32 kbps is used as the voice source.
The offered network load is defined as ρ = (λn+ λh)/(µr+ µh). To reach steady states, each simulation case is run with one million arrivals. The performance metrics are new call blocking rate, handoff call dropping rate, channel utilization, and the average voice packet dropping rate, Pd, where the channel utilization is defined as 1−BB f ree
total × 100%, not the same as the channel utilization in [72]. The quality of a voice call is considered to be acceptable if its Pd < 2%; otherwise, it is considered unacceptable.
Validation of Analytical Results
In this experiment, we assume that 40% of arrival calls are handoff calls. The channel occupancy time is 2 second. Pr is set to 0.8 when Bdeg ≤ (Btotal − Bth). Fig. 3.25 shows the blocking rate and dropping rate of both analytical and simulation results. The maximal difference of blocking rate between simulation and analytical results is about 0.44%, which appears when ρ = 20. It can be seen that analytical results match well with simulation results. So, both our analytical and simulation results are correct and believable.
0
Medium time per session (sec)
ρ
Figure 3.26: Comparisons of different schemes on: (a) channel utilization and (b) goodput.
Influence of CAC and RA
In this experiment, we want to evaluate the impact of CAC and RA. Pr is set to 0.8 when Bdeg ≤(Btotal− Bth). We compare our scheme against the CAC-only and “no-CAC-no-RA” cases. For the CAC-only case, the PIs of calls are fixed. Fig. 3.26(a) shows the channel utilization under different offered loads. Clearly, our scheme has very good utilization because calls can always be upgraded when there are extra resources. The no-CAC-no-RA (PI=20) case outperforms the CAC-only (PI=20) case because it accepts every incoming request in all network situations. Fig. 3.26(b) shows the medium time per session receives (which is approximated by the total used medium time divided by the total number of ongoing calls). With call admission control, the medium time of our scheme is better than that of the no-CAC-no-RA case. Even when the work load is high, our scheme can still guarantee the minimum bandwidth requirement of all calls. As ρ ≥ 32, the minimum time of the no-CAC-no-RA case will drop to an unacceptable level.
This shows that our scheme can well utilize network resources while guarantee the quality of calls.
Fig. 3.27 shows the average voice packet dropping rate, Pd, against different ρ for both CAC+RA and no-CAC-no-RA cases. Without CAC, the no-CAC-no-RA curve rapidly exceeds the 2% threshold, which means an unacceptable voice quality. With CAC, our algorithm has a very small Pd because the resource usage is well under control. This result shows that CAC is definitely necessary, or the quality of real-time services will easily become unacceptable.
Fig. 3.28 shows the new call blocking rate and handoff call dropping rate versus different offered loads. The rates of the no-CAC-no-RA case are all zero because every incoming request is accepted. From Fig. 3.28(a), we see that our scheme is only slightly worse than the CAC-only (PI=40) case after ρ ≥ 12 because of our call acceptance policy.
However, the benefit is our lower handoff call dropping rate, as shown in Fig. 3.28(b).
Figure 3.27: Comparison of different schemes on their average voice packet dropping rates (Pd).
Figure 3.28: Comparisons of: (a) call blocking rate and (b) call dropping rate.
Influence of Pr
The value of Pr reflects the possibility that a QAP permits new calls to start. Clearly, a larger Pr will benefit new calls but hurt handoff calls. Fig. 3.29 shows the impact of Pr on call blocking and dropping probabilities. From these curves, a suggested value of Pr could range from 0.2 to 0.6.
Influence of Traffic Characteristic
Next, we evaluate the influence of traffic characteristic. We change the percentage of handoff calls while keep the offered load unchanged. Fig. 3.30 shows this impact on call blocking and call dropping rates. We can see that our scheme is quite insensitive to this
0
Figure 3.29: The impact of Pr on: (a) call blocking rate and (b) call dropping rate.
0
Figure 3.30: The impact of the percentage of handoff calls on: (a) call blocking rate and (b) call dropping rate.
change, unless the offered load is very high. This concludes that our scheme can provide good QoS to handoff calls.
Influence of Additional BE Flows
The above experiments all assume that VoIP is the only traffic in the network. In this ex-periment, we add some additional static QSTAs, each generating best-effort data (AC BE) with a rate of 480kbps to compete with VoIP traffics. As we can see in Fig. 3.31, if AIFSN[AC BE] is too small (such as 3), the AC BE traffics will significantly affect the performance of VoIP traffics. When the voice load is high (such as ρ = 32), the value of Pd is much too high. Even in the light load case (ρ = 4), we see Pd> 2% when there are more than 20 BE streams. The effect can be reduced by enlarging the value of AIFSN[AC BE].
As shown in Fig. 3.31, with a slightly larger AIFSN for AC BE, the dropping rates for voice packets are significantly reduced. When ρ = 32 and AIFSN[AC BE]=15, Pd= 0.50%
and 0.53% for BE QSTA=2 and 8, respectively, which are improved by 89.46% and 97%, respectively, as opposed to AIFSN[AC BE]=3. This also implies that AIFSN can
effec-
AIFSN[AC_BE]
ȡ=4, BE_QSTA=2 (uplink) ȡ=4, BE_QSTA=8 (uplink) ȡ=4, BE_QSTA=20 (uplink) ȡ=4, BE_QSTA=30 (uplink) ȡ=32, BE_QSTA=2 (uplink) ȡ=32, BE_QSTA=8 (uplink) Pd(%)
Figure 3.31: Impact of interference from AC BE traffics with various AIFSNs for AC BE traffics.
tively help differentiate the priorities of voice and best-effort packets.