Resource Sharing and Bandwidth Allocation for WiMAX Mesh Networks Using Centralized Scheduling

(1)

251 Resource Sharing and Bandwidth Allocation for WiMAX Mesh Networks Using Centralized Scheduling

Resource Sharing and Bandwidth Allocation for WiMAX Mesh Networks

Using Centralized Scheduling

Yu-Liang Tang1_{, Tin-Yu Wu}2_{, Jen-Wen Ding}3_{, Jun-Jie Chen}4 1,4_{Department of Communication Engineering, Xiamen University}

2_{Department of Electrical Engineering, Tamkang University}

3_{Department of Information Management, National Kaohsiung University of Applied Sciences} 1,4_China,2.3_Taiwan

tyl@xmu.edu.cn1

Abstract

IEEE 802.16, also known as WiMAX, is considered to be an important technology for next-generation wireless metropolitan area network (WMAN) and has received great attention for its potential applications. WiMAX supports two types of modes: point-to-multipoint mode and mesh mode. While the point-to-multipoint mode is well developed, the mesh mode still has many design issues to be addressed, such as routing and packet scheduling algorithms. This paper investigates the routing and packet scheduling algorithms of WiMAX mesh networks. Three sets of routing and packet scheduling algorithms are proposed: (1) interference-based optimal routing and parallel-relay-based packet scheduling, (2) link-prioritization-based routing and link-group-based packet scheduling, and (3) minimal-conflict-cost routing and deficiency-fair priority-based packet scheduling. The simulation results show that the proposed algorithms can make efficient use of radio resource.

Keywords: WiMAX mesh networks, Route construction,

Packet scheduling, Bandwidth allocation, Interference-aware routing.

1 Introduction

In the past few years, WiMAX system has received great attention from the public for its potential applications. Compared to conventional last-mile technology like fiber-optical networks, the advantages of WiMAX include its fast deployment, easy maintenance, and low investment. WiMAX supports two modes: point-to-multipoint mode and mesh mode. Unlike IEEE 802.11a/b/g based mesh network, the 802.16-based WiMAX mesh employs TDMA to provide fine-granularity radio resource allocation, which is difficult to do so by using RTS/CTS-based radio resource allocation. Another main advantage of employing TDMA is to decrease the potential interference between neighboring nodes. There are two types of interference: primary and secondary interference. By primary interference we mean that a node performs transmission and reception at the same

time. By secondary interference we mean that a receiver R is in the transmission range of another transmitter whose destination is not R. A good wireless mesh routing algorithm should select the route with small interference. A feasible packet scheduling algorithm for wireless mesh networks should avoid interference when allocating bandwidth to the wireless nodes. In this paper, we investigate the resource sharing and bandwidth allocation problem for IEEE 802.16 WiMAX mesh networks. We propose three sets of routing and transmission scheduling algorithms for wireless mesh networks to make efficient use of radio resources.

The IEEE 802.16 mesh mode MAC supports centralized scheduling and distributed scheduling. In this paper, we adopt centralized scheduling. With centralized scheduling, every mesh subscriber station (SS) sends its resource requests to mesh base station (BS), and the mesh BS determines the amount of bandwidth allocated to each link. Route construction and packet transmission scheduling are the main design issues of wireless mesh networks. The main idea of route construction is to build a delivery tree from BS to all SS. The main idea of packet scheduling is to determine the active links for transmitting packets during each scheduling timeslot. Both design issues have great impact on network bandwidth utilization. However, it has been proven that the routing and packet scheduling problem in wireless mesh networks is an NP-Hard problem, which means that when the number of SS nodes in a wireless mesh network is large, it is infeasible to find the optimal routing and packet scheduling.

In this paper, on the basis of 802.16, we propose three sets of route construction and transmission scheduling algorithms: (1) interference-based optimal routing and parallel-relay-based packet scheduling, (2) link-prioritization-based routing and link-group-based packet scheduling, and (3) minimal-conflict-cost routing and deficiency-fair priority-based packet scheduling. The three sets of combinations not only adapt to different network conditions, but also perform better than the existing algorithms.

The rest of the paper is organized as the following. Section 2 describes the related works and the necessary

(2)

Journal of Internet Technology Volume 11 (2010) No.2

background. Section 3 discusses the proposed route construction and packet scheduling algorithms. Section 4 presents NS-2-based performance evaluation results of the proposed algorithms. Finally, Section 5 concludes this paper.

2 Related Works and Background

The IEEE 802.16 standard has explicitly brought up the notion that wireless mesh technology should be included in WiMAX, and this idea has received much attention from the enterprises because this new technology may create a new competitive least-mile technology other than DSL and cable technology. In practice, some companies have applied wireless mesh technology to broadband wireless access, and mesh-based wireless routers have been put into commercial use already. Up to now, many researchers have systematically investigated wireless access systems based on wireless mesh networks. In what follows, we briefly review recent advancements in wireless mesh networks.

In [1], the IEEE 802.16 working group elaborated the messages about 802.16 mesh networks in terms of official IEEE regulations. The algorithms regarding wireless mesh networks are proposed in [2][3][4]. Although the simulation results are not suitable for 802.16 architecture, these studies provide good points of view for further study.

In [5], Shetiya and Sharma proposed different routing and scheduling algorithms for 802.16 wireless mesh networks according to different transmission models (such as CBR and VBR). The authors use routing trees to construct the shortest paths, which is efficient in terms of network performance.

In [6], Wei et al. proposed an interference-aware timeslot scheduling algorithm. The authors quantify the potential interference for routing and define the overall blocking metric of a route as the sum of the blocking metrics of all nodes on the route. The blocking metric of a node is defined as the number of neighboring nodes that may interfere with this node. With these metrics, the path with the smallest potential interference will be selected.

In [7], Tao et al. proposed a concurrent transmission algorithm for WiMAX mesh networks. The proposed algorithm aims to transmit as many packets as possible in one minislot. Whenever a new node arrives at the network, the routing tree readjusts itself. This algorithm is the reform of the algorithm proposed in [4]. However, in terms of the total transmission time, the algorithm proposed in [7] does not yield obvious performance improvement.

In [8], Bu et al. made use of 802.16 mesh networks as the return path of 3G wireless network. In the light of IEEE 802.16 mesh topology, [9] proposed QoS mechanisms to guarantee the throughput of high-priority nodes.

Nevertheless, this study does not mention how to perform transmission scheduling to improve the overall throughput.

In [10], De Couto et al. found that in multi-hop wireless mesh networks, the shortest paths might result in poor throughput and packet delay. They suggest that in addition to bandwidth consideration, the interference issue should also be taken into account in choosing routes.

3 The Proposed Scheme

In this section, we describe the following routing and transmission scheduling algorithms: (1) interference-based optimal routing and parallel-relay-based packet scheduling, (2) link-prioritization-based routing and link-group-based packet scheduling, and (3) minimal-conflict-cost routing and deficiency-fair priority-based packet scheduling. The proposed algorithms not only suit different network situations but also consider some important factors that influence network performance, like interference, node load and delay, and make the best decision.

3.1 Interference-Based Optimal Routing

Conventional interference-aware routing algorithms construct routes in wireless mesh networks by considering the potential interference on all potential routes. For example, Wei et al. uses the number of neighbors of a node as its interference metric [6]. After calculating summing up the interference metrics of all nodes along each potential route, the route with the minimum interference will be selected. However, this algorithm has some limitation. The number of neighbors of a node is a rough metric for measuring interference. For example, consider the case shown in Figure 1, where node S1 is the source node and

node S6 is the destination node. The interference metric of

node S1 is regarded as three since it has three neighbors.

The interference metric of node S8 is regarded as one

since it has only one neighbor. However, the probability that node S1 suffers interference is not three times that of

node S8. To solve this deficiency, we use a probabilistic

metric to measure the potential interference faced by each node. Assume that the probability that a neighboring node cause interference is p. For node S1, there are tree cases

of interference: (1) only one neighboring node, S2 or S3 or

S7, creates interference; (2) two of the three neighboring

nodes create interference; and (3) all three neighboring nodes create interference. Thus, the overall probability of interference for node S1 is given by:

Pnode( )S1 p( p)2 p2( p) p3 3 1 1 3 2 1 3 3 =      − +      − +      (1) For the same reason, since nodes S2 or S4 also have

(3)

Journal of Internet Technology Volume 11 (2010) No.2

5 Conclusions

On the basis of standard IEEE 802.16, this paper first investigates the routing and transmission scheduling algorithms of IEEE 802.16 mesh network and then proposes novel routing and transmission scheduling algorithms. The proposed interference-based optimal routing is the optimization of the existing static route construction method. Furthermore, to the best of our knowledge, this paper is the first one with the attempt to propose link-prioritization-based routing algorithm and minimal-conflict-cost routing algorithm, both of which are dynamic route construction algorithms. Many important factors that have great impact on network performance, such as interference between network nodes, load on nodes, transmission delay, are all taken into account by both methods to achieve a good design tradeoff.

On the basis of the three proposed route construction algorithms, we further developed three corresponding packet scheduling algorithms, parallel-relay-based packet scheduling algorithm, link-group-based packet scheduling algorithm, and deficiency-fair priority-based packet scheduling algorithm. From the simulation results, we found that the combination of the proposed interference-based optimal routing and parallel-relay-based packet scheduling algorithm greatly outperforms the existing static route construction algorithms. Also, we found that when network nodes have the same bandwidth request, the combination of minimal-conflict-cost routing and deficiency-fair priority-based packet scheduling algorithm has the best performance in terms of network throughput, network delay, and fairness. When network nodes have different and changing bandwidth demands, the combination of link-prioritization-based routing and link-group-link-prioritization-based packet scheduling algorithm has the best performance in terms of network throughput; however, it does not perform well in terms of network delay and fairness. Overall, the combination of minimal-conflict-cost routing and deficiency-fair priority-based packet scheduling algorithm has the best performance in most of the network conditions in terms of all measured metrics.

6 Acknowledgments

Our gratitude goes to Yi-Hsin Lin and Hung-Lin Chan for their contribution to this paper.

References

[1] IEEE Standard 802.16-2004 IEEE Standard for Local and Metropolitan Area Networks-Part 16:

Air Interface for Fixed Broadband Wireless Access Systems, October, 2004.

[2] P. Hsiao, A. Hwang, H. Kung and D. Vlah, Load-balancing Routing for Wireless Mesh Networks, Proc. of IEEE INFOCOM 2001, April, 2001, pp.986-995. [3] J. Gao and L. Zhang, Load Balanced Short Path

Routing in Wireless Networks, Proc. of IEEE INFOCOM 2004, March, 2004, pp.1098-1107. [4] A. Raniwala, T.-C. Chiueh, Architecture and

Algorithms for an IEEE 802.11-Based Multi-Channel Wireless Mesh Network, Proc. of IEEE INFOCOM 2005, March, 2005, pp.2223-2234.

[5] H. Shetiya and V. Sharma, Algorithms for Routing and Centralized Scheduling to Provide QoS in IEEE 802.16 Mesh Networks, Proc. of WMuNep 2005, 2005, pp.140-149.

[6] H.-Y. Wei, S. Ganguly, R. Izmailov and Z. Haas, Interference-Aware IEEE 802.16 WiMAX Mesh Networks, Proc. of VTC Spring 2005, May, 2005, pp.3102-3106.

[7] J. Tao, F. Q. Liu, Z. H. Zeng and Z. X. Lin, Throughput Enhancement in WiMAX Mesh Networks Using Concurrent Transmission, Proc. of WCNM 2005, September, 2005, pp.871-874.

[8] T. Bu, M. C. Chan and R. Ramjee, Designing Wireless Radio Access Networks for Third Generation Cellular Networks, Proc. of IEEE INFOCOM 2005, March, 2005, pp.68-78.

[9] F. Liu, Z. Zeng, J. Tao, Q. Li and Z. Lin, Achieving QoS for IEEE 802.16 in Mesh Mode, Proc. of ICCSI 2005, July, 2005, pp.353-356.

[10] D. S. J. De Couto, D. Aguayo, B. A. Chambers and R.Morris, Performance of Multihop Wireless Networks: Shortest Path is Not Enough, ACM SIGCOMM Computer Communication Review, Vol.33, No.1, January, 2003, pp.83-88.

[11] B. Han, W. J. Jia and L. D. Lin, Performance Evaluation of Scheduling in IEEE 802.16 Based Wireless Mesh Networks, Computer Communication, Vol.30, No.4, February, 2007, pp.782-792.

[12] C. Cicconetti, I. F. Akyildiz and L. Lenzini, Bandwidth Balancing in Multi-Channel IEEE 802.16 Wireless Mesh Networks, Proc. of IEEE INFOCOM 2007, May, 2007, pp.2108-2116.

[13] D. Kim and A. Ganz, Fair and Efficient Multihop Scheduling Algorithm for IEEE 802.16 BWA Systems, Proc. of BroadNets 2005, October, 2005, pp.833-839. [14] K. Jain, J. Padhye, N. Padmanabhan and L. Qiu,

Impact of Interference on Multi-hop Wireless Network Performance, Proc. of ACM MOBICOM, September, 2003, pp.66-80.

(4)

[15] W. H. Jiao, P. Jiang, R. J. Liu and M. Li, Centralized Scheduling Tree Construction Under Multi-Channel IEEE 802.16 Mesh Networks, Proc. of IEEE GLOBECOM 2007, November, 2007, pp.4764-4768. [16] Y. B. Lu and G. Q. Zhang, Maintaining Routing

Tree in IEEE 802.16 Centralized Scheduling Mesh Networks, Proc. of ICCCN 2007, August, 2007 pp.240-245.

[17] J.El-Najjar, B.Jaumard and C.Assi, Minimizing Interference in WiMax/802.16 Based Mesh Networks with Centralized Scheduling, Proc. of IEEE GLOBECOM 2008, November, 2008, pp.1-6.

[18] Y. J. Xu, L. Z. Tao, Z. X. Fang and X. H. Min, A Bandwidth Scheduling Algorithm Based on Minimum Interference Traffic in Mesh Mode, Proc. of WiCom 2007, September, 2007, pp.822-825.

Biographies

Yu-Liang Tang currently works as an

Associate Professor in the Department of Communication Engineering, Xiamen University in China. His research interests include the next generation mobile networks and the application of mobile added-value service.

Tin-Yu Wu currently works as an Assistant

Professor in the Department of Electrical Engineering, Tamkang University, Taipei, Taiwan. He received his M.S., and Ph.D. degrees in Department of Electrical E n g i n e e r i n g , N a t i o n a l D o n g H w a Unviersity, Hualien, Taiwan in 2000, and 2007 respectively. . His research interests focus on the next generation Internet protocol, mobile computing and wireless networks.

Jen-Wen Ding currently works as an

Assistant Professor in the Department of Information Management, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan. He received his B.S., M.S., and Ph.D. degrees in Engineering Science from National Cheng Kung University, Tainan, Taiwan, in 1996, 1998, and 2001, respectively. His research interests include multimedia communications, peer-to-peer computing, and mobile computing.

Jun-Jie Chen is studying at the Department

of Communication Engineering, Xiamen University in China, for his Master Degree. His research interests in the next generation networks.