CHAPTER 4 EVALUATION RESULTS
4.1 S IMULATION E NVIRONMENT
The simulation topology is depicted in the Fig. 7. A number of MSs and a BS are connected via a gateway to a video conference endpoint and an FTP server.
Core Network
---BS
MS
Video Conference/FTP
Gateway
Video Conference
Endpoint FTP Server
Fig. 7. Simulation Topology.
The video conference application used in the simulation has variable packet size and is constrained with the latency requirement for maintaining the quality of the rtPS and FTP for the BE. The WiMAX system profile [11] and application parameters are summarized in Table 4(a) and 4(b), respectively.
Table 4. a) System profile and b) application parameters in the simulation.
(a)
System Parameter DL UL
System Bandwidth 1.5 MHz
FFT Size 1024
Sub-channels 10 12
Slots of per sub-channel 1 1
- Standard deviation: 0.75 bytes [12]
Packet inter-arrival time: 30 frames/sec FTP Requested file size: 200Kbytes
Inter-request Time: 30 sec 4.2 Simulation: Evaluation and Results
This section itemizes the scenario and criteria of evaluation focusing on the modulation-aware allocation, latency-aware dynamic adjustment, latency guarantee with different requirements and fairness.
4.2.1 Modulation-aware Allocation
Whenever the MCS is changed due to various interferences, for consistent video conferencing quality the data rate of MSs must be sustained by granting each of them adapted number of slots. Table 5 depicts the modulation awareness of the HUF, in which two MSs whose MCS is changed along with time are involved. From Figure 8 we observe that though
modulation is changed constantly, the throughput is still kept at the same rate. The fewer bits a slot carries, the more slots are granted; similar behaviors occur otherwise
Table 5. The scenario of the changed MCS in the simulation.
Modulation BPSK QPSK 16QAM 64QAM
Coding scheme 1/2 1/2 3/4 1/2 3/4 1/2 2/3 3/4
Bytes/slot 3 6 9 12 18 18 24 27
MS1 0~10 10~15 15~20 20~25 25~30 30~35 35~40 40~50 MS2 40~50 35~40 30~35 25~30 20~25 15~20 10~15 0~10
# of granted slots of MS1
# of granted slots of MS2 ThrPut of MS1
ThrPut of MS2
Fig. 8. Modulation-aware allocation: the throughput is kept whenever the MCS is changed.
4.2.2 Latency-aware Dynamic DL/UL Adjustment
The dynamic adjustment of downlink and uplink sub-frame size is used to maximize the throughput of the link. Besides, the adjustment must take the requests with latency requirement into consideration for keeping the quality. In this section, we introduce the scenario of the evaluation for the latency-aware adjustment, and compare the performances by using static adjustment and dynamic adjustment based on TPP or the proposed algorithm HUF.
Dynamic DL/UL adjustment considering the latency requirement not only maximizes the
link utilization but retains the quality of real-time applications. In this section, we evaluate the latency-aware adjustment supported by HUF and compare the performance with the static approach and the TPP. Six MSs are dedicated for downloading files using FTP with BE while an increasing number of MSs performing video conferencing with rtPS are adopted to enlarge the link load. Profiles of the applications are configured according to Table 4. Throughput and violation rate are investigated and shown in Fig. 9. Violation rate which means the ratio of the number of packets whose delay exceeds its maximum latency requirement to all number of packets is used to judge whether the adjustment is latency aware. An adjustment is said to be latency-aware if it considers the latency requirements to bring about low violation rate.
460 480 500 520 540 560 580 600
36 37 38 39 40 41
# of MSs
Throughput (kbps)
HUF TPP STA
(a)
0
Fig. 9. a) Throughput and b) violation rate of three different algorithms after DL/UL adjustment.
As depicted in Fig. 9(a), the throughput of dynamic adjustment, whether using TPP or HUF, is about 7% higher than the static adjustment when overloaded with 41 MSs. This is due to the fact that the former dynamically exploits the bandwidth according to the DL and UL traffic loads, while the latter tends to waste link resources when the traffic loads contrast much to the static allocation. Figure 9(b) shows that the degraded throughput of static adjustment contributes to the increasing violation rate. Although the TPP has similar throughput to HUF, its violation rate is considerably higher than that of HUF, whose rate is close to zero. This is because the TPP decides the DL/UL allocation simply by considering their loads, while the HUF further reserves bandwidth for requests that must be served in the current frame.
4.2.3 Latency Guarantee with Different Requirements
Latency guarantee in rtPS is critical for proper QoS. Though the requirement is different, the bandwidth allocation algorithm must guarantee and satisfy for them. In this section, we compare the proposed algorithm with the MLWDF which is throughput-optimal and considers
the waiting time of head-of-line packet to keep the latency requirement, and with the DFPQ which uses EDF [8] for rtPS to satisfy the requirement. The evaluation scenario uses the video conference application referencing Table 4(b) based on two sets of QoS parameters used in rtPS presented in Table 6. Among the parameters only one is configured differently, namely the maximum latency requirements which is 50ms and 150ms, respectively. The load of the link is increased by simultaneously increasing the input of two traffic flows.
Table 6. The QoS parameters of the two kinds of traffic flows.
QoS Parameter TYPE I TYPE II
Service Class rtPS rtPS
Minimum Reserved Rate (bps) 2400 2400
Maximum Sustain Rate (bps) 1000000 1000000
Maximum Latency (ms) 50 150
Polling Time (ms) 20 20
The criteria of the evaluation are throughput, average latency of packets and violation rate. The throughput and average latency are the general criteria to evaluate the performance of a bandwidth allocation algorithm. Besides, the evaluation scenario focuses on the satisfaction with different latency requirements, and thus takes the violation rate into account.
Latency guarantee means the violation rate is zero regardless of the requirements. Figure 10 discusses the throughput as well as the latency of three algorithms. From Fig. 10(a) we can observe that generally the throughput increases as more MSs participate in. However, the DFPQ starts to degrade when the number of MSs reaches 32. This is because the EDF, which is an optimal scheduling algorithm in resource sufficient environment, degrades rapidly when overloaded [13]. The corresponding average latency in Fig. 10(b) is thus found to exceed 1000ms suddenly from 32 MSs in DFPQ. The throughput is similar between the MLWDF and HUF, though the average latency differs by 247ms, since the MLWDF only considers the head-of-line waiting time which results in high average latency when heavily loaded. The HUF achieves high throughput while retaining low average latency.
Figure 10(c) further examines the violation of the three algorithms in latency. Even when
the number of MSs is up to 34, the HUF has no violation for the maximum latency being 50ms and 150ms. Nevertheless, the violation rate of MLWDF grows drastically when 28 MSs are involved and is close to 70% and 80%, respectively for maximum latency requirement being 150ms and 50ms when 34 MSs are present. This indicates that considering the head-of-line packet’s waiting time may not be sufficient to guaranteeing the latency requirement. The DFPQ has a violation rate of 58% for 50ms and 78% for 150ms for 34 MSs resulted from the degraded throughput.
0
0
Fig. 10. a) Throughput, b) average latency and c) violation rate of three different algorithms.
4.2.4 Fairness
A bandwidth allocation algorithm is said to be fair if the difference in normalized services received by different flows in the scheduler is bounded. [8] In point of the above description, we evaluate the fairness of the proposed algorithm HUF with DFPQ, TPP, and UPS. In the evaluation two sets of MSs are involved, in which one performs rtPS-based video conferencing and the other uploads files via BE-based FTP. The application profiles are shown in Table 4(b) while the parameters of service classes are presented in Table 7.
Table 7. The parameters of rtPS and BE.
QoS Parameter TYPE I TYPE II
Service Class rtPS BE
Minimum Reserved Rate (bps) 2400 2400
Maximum Sustain Rate (bps) 1000000 1000000
Maximum Latency (ms) 50 N/A
Polling Time (ms) 20 N/A
The fairness between rtPS and BE can be formulated as
BE
where SrtPS and ThrtPS are the requested bandwidth and the corresponding throughput of rtPS, yet SBE and ThBE are those of BE. The results are depicted in Fig. 11, in which small values suggest fair allocation. Figure 11(a) shows that TPP and HUF are fairer than DFPQ and UPS. That is because the UPS uses Strict Priority to allocate bandwidth to all service classes in which BE tends to get starved as the rtPS becomes demanding. In DFPQ, the maximum sustained rate is employed as the Deficit counter; however deciding the appropriate maximum sustained rate for all service classes is not trivial. Thus, if the maximum sustained rate is not configured properly, the fairness is degraded. Figure 11(b) further explains the results. As shown in the figure, all approaches allocate fairly, namely 17% for rtPS and 83% for BE, when 4 MSs are employed. However, UPS and DFPQ start to distribute excessive number of slots to rtPS for 8MSs because of higher priority, resulting in the starvation of BE.
Contrastively, HUF is quite fair even when 16MSs are involved. TPP behaves similarly to the HUF, but becomes much unfair when heavily loaded because it tends the proportion leads to grant more slots to one of service classes which exceed the need.
0
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
UPS TPP DFPQ HUF UPS TPP DFPQ HUF UPS TPP DFPQ HUF UPS TPP DFPQ HUF UPS TPP DFPQ HUF
4 8 12 16 20
# of MSs
Percentage (%)
BE rtPS
(b)
Fig. 11. a) Fairness and b) granted slots for rtPS and BE of four algorithms.
Chapter 5 Conclusions and Future Works
This work aims at designing an integrated bandwidth allocation algorithm for a WiMAX BS in order to guarantee the latency requirement of real-time applications as well as service differentiation and fairness among all service classes. Dynamic downlink/uplink adjustment is also employed to well utilize the scarce wireless link. Since the mobility is supported in the WiMAX system in which link quality frequently changes due to long distance and interference, the modulation and coding schemes need to be adaptive to the link status between MSs and BS. Moreover, the Grant-Per-SS (GPSS) is preferred not only to comply with the standard but also to provide MSs the flexibility of domination.
The Highest Urgency First (HUF) is proposed to achieve the above goals. HUF translates the requested size to number of slots according to current MCS when every frame starts and uses the Urgency of data/request as criterion of allocation. A data/request with a deadline equaling to one is the most urgent and needs to be allocated immediately, while others’ urgency is decided by the U-factor. In the dynamic DL/UL sub-frame determination, the HUF firstly reserves bandwidth for (1) data/requests whose deadline equals to one and (2) the minimum reserved rate of each service flow, and then proportionates the remaining bandwidth for DL/UL according to the remnant non-urgent data/requests. After satisfying (1) and (2) for DL/UL allocation to queues, the head-of-line data/request of a queue with the largest average-U-factor is granted one by one until the sub-frame is fulfilled. Finally, each MSs obtains its grant from its own service queues.
Simulation result indicates that the quality is retained as the MCS adapts owing to the link quality. For dynamic adjustment, we show the throughput is good as TPP and increases 7%
compared with static adjustment, and the violation rate is better about 42% and 80% than TPP and static adjustment respectively. HUF outperforms the DFPQ by 25% in throughput when
overloaded, and incurs no latency violation within system capacity. Finally, we compare the fairness of UPS, DFPQ, TPP and HUF and observe fairness between rtPS and BE in HUF which avoids inappropriate grant for rtPS.
Though HUF is relatively tolerant to overloaded situations, as a future direction, we plan to develop admission control schemes to ease the degradation of the throughput and fairness.
Besides, while latency guarantee and fairness are now concerned in BSs, bandwidth allocation algorithms for MSs are also demanded to schedule appropriately the granted bandwidth.
Reference
[1] IEEE 802.16 Working Group, “Air interface for fixed broadband wireless access systems,”
Jun 2004.
[2] Cable Television Laboratories Inc., “Data-Over-Cable Service Interface Specifications - Radio Frequency Interface Specification v1.1,” July 1999.
[3] IEEE 802.16 Working Group, “Air Interface for Fixed and Mobile Broadband Wireless Access Systems - Amendment for Physical and Medium Access Control Layers for Combined Fixed and Mobile Operation in Licensed Bands,” Feb. 2006.
[4] IEEE 802.11 Working Group, “Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications,” Sep. 1999.
[5] W. M. Yin, C. J. Wu, Y. D. Lin, “Two-phase Minislot Scheduling Algorithm for HFC QoS Services Provisioning,” GLOBECOM, Nov. 2001.
[6] M. Andrews et al., “Providing Quality of Services over a Shared Wireless link,” IEEE Communication Magazine, pp. 150-154, Feb. 2001.
[7] K. Wongthavarawat, A. Ganz, “IEEE 802.16 Based Last Mile Broadband Wireless Military Networks with Quality of Service Support,” MILCOM, Oct. 2003.
[8] J. Chen, W. Jiao, H. Wang, “A Service Flow Management Strategy for IEEE802.16 Broadband Wireless Access Systems in TDD Mode,” ICC, May 2005.
[9] Y. N. Lin, S. H. Chien, Y. D. Lin, Y. C. Lai, M. Liu, "Dynamic Bandwidth Allocation for 802.16e-2005 MAC," Book Chapter of "Current Technology Developments of WiMax Systems," edited by Maode Ma, to be published by Springer, 2007.
[10] A. Sayenko, O. Alanen, J. Karhula, T. Hamalainen, ”Ensuring the QoS Requirements in 802.16 scheduling,” ACM MSWiM ’06, Oct. 2006.
[11] Mobile WiMAX Part I, “A Technical Overview and Performance Evaluation,” WiMAX Forum, April 2006.
[12] D. P. Heyman, A. Tabatabai, T. V. Lakshman, “Statical Analysis and Simulation Study of Video Teleconference Traffic in ATM Networks,” IEEE Trans. Circuits Sys. Video Tech., vol. 2, no. 1, Mar. 1992, pp.49-59.
[13] C. Lu, J. A. Stankovic, G. Tao, S. H. Son, “Design and Evaluation of a Feedback Control EDF Scheduling Algorithm,” Real-Time Systems Symposium, 1999.