ISSUES ON QOS SCHEMES IN WIRELESS SENSOR NETWORKS

. UFMG - ICEX DEPARTAMENTO DE CIÊNCIA DA C O M P U T A Ç Ã O UNIVERSIDADE FEDERAL DE MINAS GERAIS ISSUES ON QOS SCHEMES IN WIRELESS SENSOR NETWORKS RT.DCC.004/2005 LUIZ H. A. CORREIA DANIEL F. MACEDO ALDRI L. DOS SANTOS JOSÉ M. S. NOGUEIRA ABRIL 2005

Issues on QoS Schemes in Wireless Sensor Networks Luiz H. A. Correia 1,2, Daniel F. Macedo 1, Aldri L. dos Santos 1, José M. Nogueira 1 1 Dept. of Computer Science 2 Dept. of Computer Science Federal University of Minas Gerais, Brazil Federal University of Lavras, Brazil Abstract Quality of Service (QoS) is an important requirement for the well-functioning of traditional and new networks. Wireless Sensor Networks (WSN) are prone to numerous events due to their mobility, harsh communication medium, and environment behavior. Little attention has been paid to the QoS requirements of such networks. This work presents a preliminary study on the use of QoS in WSN. We focus our research on QoS techniques applied to medium access control (MAC), and propose a classification of such techniques in four categories. 1 Introduction Wireless Sensor Networks (WSN) are a novel type of wireless network where hundreds or thousands of energy-constrained devices cooperate to monitor and/or actuate in a given area. Those networks allow real-time monitoring of regions where the presence of humans may be restricted or inappropriate, such as disaster sites, dense forests and radioactive plants. In some applications, such as nuclear plant supervision, the network must provide quality of service (QoS) guarantees to the data being collected. Due to memory, bandwidth and processor limitations, QoS in WSN is a challenging task. Such QoS policies must be kept simple yet effective. The design of QoS policies for WSN is also hardened as the network runs one big, distributed application. In traditional networks, QoS focus on application differentiation, e.g. delay-sensitive traffic over delay-insensitive traffic. In WSN, QoS mechanisms focus on priorization (stablishing an order) among concurrent tasks of a given application. Thus, due to the highly specific requirements, QoS techniques must allow applications to dictate their own QoS policies. Currently, the literature pays little attention to QoS issues in WSN. In this article we present an overview of QoS issues in networking protocols, more specifically in the medium access control (MAC) layer. Our aim is to contribute to the design of QoS-aware MAC protocols, by providing a categorization and examples of QoS techniques that could be used in sensor networks. Those categories are derived from mechanisms employed in the MAC layer in WSN and traditional networks. Partially supported by CNPq, Brazil, process number 55.2111/2002-3. Issues on QoS Schemes in Wireless Sensor Networks - DCC/UFMG - Technical Report: RT.DCC.004/2005 Pág. 1

The rest of paper is organized as follows. Section 2 presents some QoS techniques recently applied in wireless sensor networks. Section 3 introduces a separation of those techniques focused on their application on MAC for WSNs. Section 4 presents the final considerations. 2 QoS techniques in WSNs Traditionally, QoS is dealt as the priorization of a class of transmissions/application over others. In WSN, as nodes cooperate to perform a single task, the wireless medium is employed to coordinate node activities and transmit data collected to a Base Station. Thus, QoS in WSN implements policies for priorization of specific application functions. Next, we briefly mention some QoS techniques employed in each layer of the protocol stack. Physical layer: this layer seeks to avoid the interference with other networks or natural sources of radiation, employing techniques that modify the modulation and frequency of operation in order to enhance the signal to noise ratio (SNR), as in HiperLAN [8], as well as HomePlug and HomePNA [7]. In WSNs, due to cost and energy constraints, modulation and frequency are fixed to simplify the radio. Link layer: in this layer the scheduling of medium access and the sequence of packets to be sent are changed to supply the QoS requirements. These changes may be achieved by packet reordering and by priority control and admission policies. It is possible to adjust the amount of control packets sent to increase the QoS of a data packet. Those techniques will be detailed in Section 3. Network layer: similar to the link layer, it is possible to use packet priorization. In wired networks, packet priority is assigned by flow or individually, using packet queuing policies such as token bus and WFQ (Weighted Fair Queuing), among others [5]. In WSN, in general, there is not the notion of a flow, precluding per flow priorization. Due to the variability of link quality, both the amount of copies sent and the number of distinct routes for a given packet can be adjusted according to its priority, increasing the probability of a successful delivery of a packet [1]. Application layer: in wired networks, audio and video applications employ adaptive compression techniques which adjust to network conditions in order to reduce the bandwidth used [5]; the same approach can be applied to sensor networks, adjusting the frequency in which data are sent [6]. Due to the intrinsic characteristics of sensor networks, the design of QoS-enabled protocols for WSN differ from traditional approaches.in traditional networks, QoS is based on metrics such as bandwidth, latency and jitter. In WSN, we face various challenges that are not found in traditional networks, mentioned below: Scarce energy: in order to prolong the lifetime of a WSN, protocols need to consider the residual energy of the nodes. Hence, QoS metrics in sensor nodes must also take into account the energy consumed in the transmission of data [11]. Data correlation: data collected by a sensor node tends to be correlated to the data collected by its neighbors, thus it may be dropped [6] or submitted to fusion or preprocessing in order to decrease bandwidth utilization [11]. Low power operation: sensor nodes permute operational and non-operational periods to increase energy Issues on QoS Schemes in Wireless Sensor Networks - DCC/UFMG - Technical Report: RT.DCC.004/2005 Pág. 2

savings [3]. Hence, QoS protocols must dynamically adapt to the state of nodes, providing uninterrupted and predictable service even as nodes are turned on/off. Hardware restrictions: the resource limitation imposed to the sensor nodes constrains the development of QoS policies toward simple solutions. Those solutions must have few memory and processing costs, and must be bandwidth-efficient. Noisy medium: high bit error rates in the communication channel are characteristic of wireless networks, and must be pondered in the design of QoS policies. It affects the number of retransmissions and the delivery rate. Further, the bit error rate may impact which validation mechanism will be adopted on the transmission, according to the importance of the data: CRC (Cyclic Redundancy Check) or FEC (Forwarding equivalence Class). In view of factors mentioned above, Younis et al. proposed the addition of two new QoS parameters: residual energy and packet delivery rates [11]. Sankarasubramanian et al. argue that QoS policies should instead guarantee freshness and accuracy properties on the current sensed area knowledge, i.e. the difference between the current information gathered by an outside observer in relation to the current state of the sensed area [6]. Due to the urgency of a QoS proposal for WSNs we suggest the junction of QoS techniques used in both wired networks and wireless networks to current proposals in WSN. We believe that various solutions employed in traditional networks can be adapted to WSN. 3 QoS techniques in MAC protocols In this section, we propose the classification of MAC layer QoS techniques for WSN in four categories: packet reordering, medium access priorization and allocation, admission control, and control overhead. Packet reordering: this category of algorithms employs outgoing packet reordering to adjust the time spent by the packets in queue according to their priority. The IEEE 802.11e standard apply this technique separating the packet queue in priority classes, which are defined by the higher layers [10]. RAP (Real- Time Architecture Protocol) employs packet reordering in WSNs [4]. Since sensor nodes have severe resource restrictions, packets will be divided among a few number of classes. Medium access priorization and allocation: this category of algorithms employs priorization policies that increase the probability of nodes with outgoing high priority packets to access the medium. In contentionbased protocols (such as 802.11e in EDCF Enhanced Distributed Coordination Function mode [10]), differentiated backoff policies and wait times are employed for each packet priority. In contention-free protocols, such as the IEEE 802.15.4 in superframe mode, slots are reserved for high priority packets [2]. WSN may use both mechanisms to enforce latency requirements. Admission control: this category of algorithms control the amount of data in which a node may send in a given time period. Admission control is used in 802.11e, for example, where the base station may determine how many bytes a given station can send in each superframe [10]. In WSN, admission control will probably be implemented as distributed rate-adaptive MAC protocols, which signal to the application whenever it may increase/decrease the frequency of packets sent (as in ARC [9]). Control overhead: these algorithms selectively adjust the amount of control packets used for each data Issues on QoS Schemes in Wireless Sensor Networks - DCC/UFMG - Technical Report: RT.DCC.004/2005 Pág. 3

packet in order to guarantee higher delivery rates. This mechanism increases delivery rates at the expense of lower bandwidth and higher energy consumption. Thus, according to the priority of the packet, acknowledgment frames and RTS/CTS dialogs can be turned on/off. Further, transmissions which demand no losses, such as reprogramming tasks, would use control packets on every packet [1]. 4 Conclusion Quality of Service is a crucial aspect in the design of WSN. Information sent by the nodes must be delivered according to QoS requirements for providing to outside observers a timely and accurate vision of the monitored phenomena. In this article we presented the key issues faced in the design of QoS protocols, namely MAC protocols. We presented a categorized overview of available QoS techniques passive to be employed in a QoS MAC protocol, and described how those techniques can be implemented. References [1] S. Bhatnagar, B. Deb, and B. Nath. Service differentiation in sensor networks. In Proceedings of the Fourth International Symposium on Wireless Personal Multimedia Communications, September 2001. [2] E. Callaway, P. Gorday, L. Hester, J.A. Gutierrez, M. Naeve, B. Heile, and V. Bahl. Home networking with IEEE 802.15.4: a developing standard for low-rate wireless personal area networks. IEEE Communications Magazine, 40(8):70 77, August 2002. [3] F. Dai and J. Wu. Energy-Efficient Coverage Problems in Wireless Ad Hoc Sensor Networks. Journal of Computer Communications on Sensor Networks, 2004. [4] Chenyang Lu, Brian M. Blum, Tarek F. Abdelzaher, John A. Stankovic, and Tian He. RAP: A real-time communication architecture for large-scale wireless sensor networks. In Proceedings of the Eighth IEEE Real- Time and Embedded Technology and Applications Symposium, page 55. IEEE Computer Society, 2002. [5] Larry L. Peterson and Bruce S. Davie. Computer Networks: A Systems Approach, 3rd Edition. Morgan Kaufmann Publishers Inc., 2003. [6] Yogesh Sankarasubramaniam, Özgür B. Akan, and Ian F. Akyildiz. ESRT: event-to-sink reliable transport in wireless sensor networks. In Proceedings of the 4th ACM international symposium on Mobile ad hoc networking & computing, pages 177 188. ACM Press, 2003. [7] Pedro Velloso, Daniel Cunha, Aurelio Amodei Junior, Marcelo Rubinstein, and Otto Carlos Muniz Bandeira Duarte. Redes Domiciliares: Princípios e Desafios das Tecnologias sem Novos Fios. In 22nd Simpósio Brasileiro de Redes de Computadores, pages 221 268, Maio 2004. [8] Bernhard Walke, Norbert Esseling, Jörg Habetha, Andreas Hettich, Arndt Kadelka, Stefan Mangold, Jörg Peetz, and Ulrich Vornefeld. IP over Wireless Mobile ATM - Guaranteed Wireless QoS by HiperLAN/2. Proceedings of the IEEE, 89:21 40, Jan 2001. [9] Alec Woo and David E. Culler. A transmission control scheme for media access in sensor networks. In Mobile Computing and Networking, pages 221 235, 2001. [10] Yang Xiao. IEEE 802.11e: A QoS Provisioning at the MAC layer. IEEE Wireless Communications, 11(3):72 79, Jun 2004. [11] Mohamed Younis, Kemal Akkaya, Mohamed Eltoweissy, and Ashraf Wadaa. On handling qos traffic in wireless sensor networks. In Proceedings of the 37th Annual Hawaii International Conference on System Sciences - Track 9, page 90292.1. IEEE Computer Society, 2004. Issues on QoS Schemes in Wireless Sensor Networks - DCC/UFMG - Technical Report: RT.DCC.004/2005 Pág. 4