摘要
随着人们对个人通信需求的日益增长,以及一些特殊工作场合的要求,传统的具有基础设施的无线网络显得有些无能为力,于是移动自组织网Ad Hoc网络便迅速发展起来。但是,Ad Hoc网络的移动自组织性、网络动态拓扑以及有限带宽等特点,严重制约了它向用户提供的服务质量。因此,对Ad Hoc网络的QoS研究是具有很重要的理论和实际意义的。
媒体接入控制MAC层处在Ad Hoc网络协议栈的底层,它通过提供有效的调度机制来分配有限的信道资源,直接控制着所有数据报文和控制信息在无线信道上发送和接收。因此,MAC层的调度机制是上层各种协议机制提供QoS的最终保障因素。通过对MAC层接入控制的研究来提供QoS保障已经成为研究热点。
本文对Ad Hoc网络MAC层的接入控制机制进行深入分析后,提出了基于窗口阶的折半退避算法—ZDCF,之后结合QoS保障的要求,将ZDCF算法应用到了增强型的MAC接入控制机制EDCA中,提出了通过检测网络状态动态修改竞争机会(TXOP)值的策略—ZEDCA。最后在Linux操作系统中建立NS2仿真模型,实现对MAC层接入控制的模拟,分析实验产生的trace结果后,得出本文提出的改进机制能够很好的保证QoS要求。
With the people's growing demand for personal communications, and the rapid emergence of wireless communications technology and internet, wireless communication network has been developed rapidly in recent years. Because of some special occasions and life requirements, traditional infrastructure-based wireless network appeared to be quite helpless, at this time Mobile Ad Hoc Network (Ad Hoc) became the research hotspot. In the Ad hoc network, each node has both router’s and host’s functions. It can be widely applied to military communications, post-disaster rescue and business occasions. But meanwhile, Ad Hoc network has the characteristics of auto-formed, dynamic topology and distributed which seriously restricted its service providing to users, therefore Ad Hoc network QoS research has important theoretical and practical significance. Media access control (MAC) sub-layer is at the bottom of Ad Hoc network’s protocol stack, its role is to provide an effective mechanism for allocating limited wireless channel resources, therefore, whether the MAC layer can efficiently use the limited wireless channel bandwidth can be a key factor to the protection of QoS provided by the upper protocol mechanisms. The characteristic of wireless channel is mainly embodied in the following layers of network layer, and some QoS mechanism can be applied more efficiently in the MAC layer, so it is also a hotspot to research the QoS protection in the MAC layer. Ad Hoc Network MAC layer protocol can be divided into random competition category, type and mixed type of distribution protocol, the current studies mainly used IEEE 802.11 MAC protocol which has been most widely used. Therefore, this paper will also use the IEEE 802.11 MAC protocol in Ad Hoc Network MAC Protocol for QoS research.
In this paper, the location, the structure and the frame format of the MAC layer are introduced firstly, and then by analyzing the distributed coordination function (DCF) mechanism and its basic back-off algorithm BEB, an exponential-based half back-off algorithm ZDCF is proposed. Besides, combining the needs of QoS protection, ZDCF algorithm is applied to the low-priority business window back on mechanism of DCF Enhanced Distributed Channel Access (EDCA), and a new strategy of ZEDCA is proposed which dynamically modifies Transmission Opportunity (TXOP) by detecting the status of network.
DCF is the lower sub-layer of the MAC layer which uses a competitive algorithm to provide mechanism of access for all communications of distributed network, and it is widely used in the Ad Hoc network. The core mechanism of DCF is Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA) protocol, which has two strategies including the basic operational mode and the RTS/CTS mode. In the DCF mechanism, if collision in channel happens while sending data, the process of competition back window is needed, the window of CWmin and CWmax is decided by the physical layer, DCF uses Binary Exponential Back-off (BEB). In BEB, if the collision of data has happened, the back window will be doubled, and after the data is sent successfully, the window will back to its smallest size CWmin, but data congestion will inevitably take place at the smallest. In this paper, by analyzing various improved schemes of BEB algorithm, an exponential-based half back-off algorithm ZDCF is proposed. ZDCF uses BEB scheme algorithm while the collision of sending data happens, however, doubles the window, and after the data has been sent successfully, the mid value of the current window order and the minimum window order is got as the order of the back-off window instead of backing straight to the minimum window. The ZDCF algorithm only needs to add a variable in the program to store the current order of the window, which belongs to the static change of CW algorithm, in line with the needs that because of the wireless channel bandwidth limitations and the characteristics of instability, it shouldn’t transmits control information frequently. While there aren’t many nodes, the window will back to the minimum value immediately, so it has the advantage of BEB algorithm; while there are many nodes, the back window will be located in different orders after calculating and won’t cause congestion, and it avoids the drawback of BEB algorithm. ZDCF computing approach is simple, not easy to make mistakes, and better scalability, meanwhile, it can be used in different window scope without any changes.
To enhance the function, IEEE Standards Committee Working Group sets up the E working group. By the end of 2005, it completed the E standard-setting work, and released 802.11e protocol, provided the reference standard for the implementation of QoS. The competed distribution mechanism in 802.11e is EDCA, which is the DCF enhanced QoS control. EDCA has set up four different priority access queue AC and set different AIFS, CWmin, CWmax, and TXOP for different AC, the smaller the values are, the higher priority the business of corresponding AC are. As can be seen through the analysis, the low priority business has a larger scope window in EDCA, and after the data has been sent successfully, the window backs off to the minimum window, so it has the drawback of BEB. In this paper, the ZDCF algorithm is applied to the low priority business of EDCA, without prejudice of the EDCA setting, the low priority business is improved. TXOP is the introduction of the new EDCA parameters, and it allows the sites to send data continuously without channel competition during TXOPlimit, so the larger the TXOP is, the higher priority of AC is. By analyzing the results of some research we found that, when there is a heavy network load, increases the TXOP can enlarge the system’s throughput. When there is a light network load, TXOP can be adjusted lower to avoid the“starve to death”condition of low priority business. So, this paper presents a strategy which dynamically changes TXOP by the detection of network status, it considers the current network status by the variable block_count which is the amount of sending data’s collision and retransmit, the large block_count means a heavy network load. Besides, consideration based on sound, ZEDCA also introduced a transmission factor and a benchmark value BLOCK to control the scope of calculation results in order to avoid shaking seriously. In ZEDCA mechanism, the formula which calculates TXOP value can make TXOP positively change with the load of network, and under certain circumstances improves the performance of the system.
Finally, a NS2 simulation model is set on Linux operating system and the real-time system simulation results is shown by the NAM network animation, then the data stored in the trace file which is generated by the simulation is analyzed. The analysis result proves the rationality of the theory proposed in the paper that it can guarantee the QoS.
引文
[1]郑少仁,王海涛,赵志峰等. Ad Hoc网络技术[M].北京:人民邮电出版社, 2005: 1-292.
[2]郑相全等.无线自组织网使用教程[M].北京:清华大学出版社, 2004: 1-380.
[3]司鹏博,胡亚辉,程源等.无线宽带接入新技术[M].北京:机械工业出版社, 2007: 240-266.
[4] Chakrabarti S. QoS issues in Ad hoc Wireless Net-works[J]. IEEE Communications Magazine, 2001, 39(2): 142-148.
[5]汪涛.无线网络技术导论[M].北京:清华大学出版社, 2008: 157-178.
[6] IEEE Std 802.11-1999 (R2003): Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications. Jun. 12, 2003.
[7] IEEE Std 802.11e-2005: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications Amendment 8: Medium Access Control (MAC) Quality of Service Enhancements, 2005.
[8]于斌,孙斌,温暖,王绘丽,陈江锋. NS2与网络模拟[M].北京:人民邮电出版社, 2007: 116-169.
[9] William Stallings著,何军等译.无线通信与网络[M].北京:清华大学出版社, 2005: 335-352.
[10]刘化君等编著.计算机网络原理与技术[M].北京:电子工业出版社[M], 2005: 117-120.
[11]曾华桑.现代网络通信技术[M].成都:西南交通大学出版社, 2004.
[12] E.Ziouva, T.Antonakopoulos. CSMA/CA performance under high traffic conditions: Throughout and delay analysis[C], Comp. Commune, vol. 25, 2002: 313-321.
[13] G.Bianchi, Performance Analysis of the IEEE 802.11 Distributed Coordination Function [J]. IEEE Journal on Selected Areas in Commu, 2000, 18(3): pp.535-547.
[14] Wireless Medium Access Control(MAC) and Physical Layer(PHY) Specifications. Medium Access Control(MAC) Enhancements for Quality of Service(QoS), IEEE std 802.11e/D2.0: Draft Supplement to IEEE Std 802.11, 1999 Edition[M].[S.L.]:[s.n.], 2001.
[15] F.Cali, M.Conti, E.Gregori, IEEE 802.11 protocol: design and performance evaluation of an adaptive backoff mechanism [J], IEEE Journal on Selected Areas in Communications, 2000, 18(9): 1774-1786.
[16] G.Bianchi, I.Tinnirello, Kalman filter estimation of the number of competing terminals in an IEEE 802.11 network [C] // IEEE, Twenty-Second Annual Joint Conference of the IEEE Computer and Communications Societies. America: IEEE, 2003: 844-852.
[17] Bharghavan V, Demaers A, and Shenker S. MACAW: A media access protocol forwireless LANs[C]. Proceedings of ACM SIGCOMM’94. London, UK, 1994: 212-225.
[18] WU H, CHENG S, PENG Y, et al. IEEE 802.11 distributed coordination function (DCF): analysis and enhancement [C] // In Proceedings of ICC’02. New York: [s.n.], 2002: 605-609.
[19] Qiang Ni, Imad Aad, Chadi Barakat and Thierry Turletti.“Modeling and analysis of slow CW decrease of IEEE 802.11 WLAN”[J]. In Proceedings of PIMRC. 2003.
[20] Song N O, Kwak B J, Song J. Enhancement of IEEE 802.11 distributed coordination function with exponential increase exponential decrease backoff algorithm[C]. The 57th IEEE Semiannual,Orlando, Florida: Vehicular Technology Conference, 2003: 2775-2778.
[21] Qixiang Pang, Soung C. Liew, Jack Y. B. Lee, et al. Performance evaluation of an adaptive back-off scheme for WLAN[J]. Wireless Communications and Mobile Computing, 2004, 4(8): 867-879.
[22] IEEE. IEEE Standard for Wireless LAN Medium Access Control(MAC) and Physical Layer (PHY) Specifications: Amendment 7: Medium Access Control(MAC) Quality of Service (QoS) Enhancements Draft 9.0, 2004.
[23] IEEE 802.11e / D7.0, Wireless LAN Medium Access Control(MAC) and Physical Layer (PHY) Specifications: Medium Access control (MAC) Enhancements for Quality of Service (QoS)[S], 2004.
[24]阮加勇,黄本雄等. IEEE 802.11e EDCF性能评估[J].华中科技大学学报, 2004, 32(6): 7-9.
[25]伍智超,王婷,贺昕,周正. IEEE 802.11e EDCA和CFB下饱和吞吐量分析[J].信息传输与接入技术, 2006, 32(6).
[26]黄志昊,李珊君. IEEE 802.11e EDCA在CFB模式下的性能分析[J].电子测量技术, 2007, 30(5).
[27]秦冀,姜雪松.移动IP技术与NS-2模拟[M].北京:机械工业出版社, 2006: 85-213.
[28]徐雷鸣,庞博,赵耀. NS与网络模拟[M].北京:人民邮电出版社, 2003: 108-125.
