Statistical Multipath Queue-Wise Preemption Routing for ZigBee-Based WSN
- 397 Downloads
Nowadays, wireless sensor network (WSN) is an important component in IoT environment, which enables efficient data collection and transmission. Since WSN consists of a large number of sensor nodes, network congestion can easily occur which significantly degrades the performance of entire network. In this paper a novel scheme called SMQP (Statistical Multipath Queue-wise Preemption) routing is proposed to balance the load and avoid the congestion for ZigBee-based WSN. This is achieved by employing statistical path scheduling and queue-wise preemption with multiple paths between any source and destination node. NS2 simulation reveals that the proposed scheme significantly improves the QoS in terms of delivery ratio, end-to-end delay, and packet delivery ratio compared to the representative routing schemes for WSN such as ad hoc on-demand distance vector and ad hoc on-demand multipath distance vector scheme.
KeywordsCongestion Wireless sensor network Multipath routing Load balancing Packet scheduling Real-time Non-real-time
Wireless sensor network (WSN) consists of a number of sensor nodes, and nowadays it is widely used in various application domains. WSN is imperative for timely and accurate detection of critical events in any target area especially when the access is restricted. Unlike wired network, it has limitations in network lifetime and communication range between the nodes. ZigBee is usually employed for WSN , and it is a high-level communication protocol based on IEEE 802.15.4 standard specifications. With several favorable features of ZigBee protocol including low-power and secure networking, it has been the best choice for various environments such as IoT, home automation, health care, smart energy, etc. . For IoT, ZigBee has been further improved with respect to data routing and scheduling . Here different types of packets are sent with different priority while fairness is considered. For this, efficient packet scheduling is needed in each node of the WSN.
In WSN congestion occurs when many nodes simultaneously transmit data, which significantly deteriorates the network performance. They increase queuing delay, packet loss, and the blocking of new connections. When some nodes or links become overloaded due to congestion, transmission failure occurs and data are lost. Network congestion or bottleneck can be avoided through multipath routing and load balancing. Multipath routing is the technique distributing the network traffic on all available paths, and thus balancing the traffic between the source and destination . With ZigBee mesh topology, a node may have several paths for a packet leading to the destination. Enhanced bandwidth and reliability are the main benefits of multipath routing technique. Ad hoc on-demand multipath distance vector (AOMDV) [5, 6] is an extended version of ad hoc on-demand distance vector (AODV)  routing protocol which not only emphasizes optimal short path routing but also considers load balancing and distribution of network traffic via multiple paths. It provides equivalent paths of the same hop counts to the IoT gateway. It can improve the reliability of the network, reduce the number of overloaded nodes, and avoid link congestion. The paths must be disjoint, however, which unnecessarily limits the number of available alternative paths. Besides, it does not consider simultaneous usage of the paths.
In WSN the First Come First Serve (FCFS) policy  is typically employed for packet scheduling, where the packets are served by the order of arrival. In some situation, however, urgent packets need to be quickly delivered to the destination regardless of the order. Here the priority is decided based on various parameters of the packet such as deadline, size, and type, etc. Numerous studies on packet scheduling for WSN have been reported in the literature along with the planning of sleep–wake time [8, 9, 10, 11, 12, 13, 14]. However, most existing packet scheduling algorithms of WSN do not reflect the urgency of the packet in the scheduling [15, 16]. Even though extensive researches have been made on multipath routing, a single path of minimum cost is usually recommended among several paths while the secondary path is taken only when the primary path fails. An efficient and dynamic packet scheduler algorithm needs to be developed to prioritize the packets based on the urgency. Here the intermediate nodes require to change the order of delivery of the packets waiting in the queue based on the priority. Preemption [17, 18] is usually employed as an OS utility for the tasks running on a system. It is an effective approach for avoiding the delay of high priority tasks while allowing the system to be fairly shared among regular tasks. The preemption mechanism needs to be efficient and lightweight so that it can react fast to the dynamic system load.
Statistical path scheduling and queue-wise preemption for maximizing the bandwidth utilization
Multipath routing for different type packets for supporting high efficiency and QoS
Generic queuing model applicable to the analysis of packet control scheme for WSN
2 Related Work
In this section the work related to packet scheduling for WSN-based ZigBee protocol is presented. The packet forwarding scheme in Reliable Real-Time Protocol (R2TP)  is based on the time metric, where the main focus is to achieve reliable transmission by duplicating the packets and using multiple paths. The velocity monotonic scheduling (VMS)  considers the end-to-end deadline of each packet before transferring a packet toward the destination by reflecting the local urgency. Compared with non-prioritized packet scheduling technique, VMS reduces the packet drop ratio by giving higher priority to the packets requesting higher velocity. Here data are prioritized based on the distance from the source node to the base station (BS) and deadline. If the deadline of a particular task expires, it is dropped at an intermediate node.
The Priority-based Congestion Control Protocol (PCCP) scheme  prioritizes both source and transit traffic, with some restriction in the process of sensed data in each node. The RACE scheme  follows the Bellman–Ford algorithm to find the paths between the source and destination showing minimum traffic load and delay. It decides the priority of a packet using the earliest deadline first (EDF) policy, and uses a prioritized MAC protocol. When the channel becomes idle, it modifies the initial wait time. The Adaptive Double Ring Scheduling (ADRS) scheme  divides the traffic into two queues according to the priority decided by the EDF policy. The scheduler dynamically switches between two queues based on the deadline of newly arrived packets.
In  the number of queues of a node is decided according to its location in the network. Since the packet generated from the node far from the BS usually takes relatively long time to reach the BS, the order of the delivery of the packets is changed in the intermediate nodes. For this, the queue and the position inside the queue for each packet are decided according to the priority. In the scheduling algorithm for WMSN , the buffer space of intermediate nodes is divided into four queues to hold three different types of video frames and one regular data frames. The video frames are given higher priority with the scheduling in round-robin fashion. Data frames are transmitted in FCFS discipline when the first three queues are empty.
Considering the types (priorities) of typical packets of WSN, the proposed scheme classifies them into three types; control, real-time, and regular packet. This will let the control packets traverse the network most swiftly while significantly reducing the end-to-end delay of real-time packets. The proposed scheme is presented next.
3 The Proposed Scheme
In this section the proposed scheme for ZigBee-based WSN is presented. It prioritizes the packets to reduce the drop rate of high priority packets, and effectively utilizes the resource of the network and thus minimizes the average waiting time and end-to-end delay. It also tries to avoid congestion by employing multipath routing which in turn increases the delivery ratio of the packets.
3.1 Design Goals
The design goals of the proposed scheme are (1) prioritization of the packets based on the type, (2) minimization of the drop ratio of control and real-time packets against regular type packets, (3) employment of multiple queues with statistical scheduling to avoid congestion and minimize energy consumption while ensuring high delivery ratio of packets under the fairness policy.
In wireless and wired communication, the end-to-end delay of packet transmission is regarded as one of the important quality metrics, especially for real-time applications. If buffer overflow and congestion occur in the network, packet loss rate is increased. Also, network throughput and energy efficiency of sensor nodes are decreased. In order to decrease packet loss rate and queuing delay, network congestion needs to be avoided.
In the proposed scheme three queues, Q1, Q2, and Q3, of different queuing discipline are used for accommodating the packets of each of three priorities, respectively. Each of them is associated with a path such as Q1 with P1, Q2 with P2, and Q3 with P3. Here statistical path scheduling is applied such that if Q1 is empty, the packets in Q2 are distributed to P1 and P2. Similarly, if both Q1 and Q2 are empty, the packets in Q3 are distributed to all the three paths. With the proposed approach, different types of packets are forwarded to the receiver simultaneously on multiple paths, maximizing the path utilization. Delivery ratio, end-to-end delay, and packet drop ratio are the performance metrics taken to evaluate the proposed scheme.
3.2 Packet Scheduling
3.3 Performance Modeling
The variables of used in the model
Mean arrival rate of the packet of priority_i (i = 1,2,3)
Mean service rate of the packet of priority_i (i = 1,2,3)
Wait time of a packet in the queue
Delay time of a packet in a node
Number of packets in queue_Q
Probability of a node having n packets
Using the model of the target WSN, the values of the network control parameters are decided to avoid congestion and increase the QoS of the network. The key parameters are the traffic generation rate and service rate in each node, packet scheduling rate, etc.
4 Performance Evaluation
The parameters used in the simulation
Number of nodes
50 × 50 m2
Number of packets sent
Physical and MAC protocol
CBR (constant bit rate)
Link layer type
The comparison of the three routing schemes
Packet delivery ratio (%)
In this paper a novel packet scheduling and routing scheme has been proposed for ZigBee-based WSN. It effectively reduces congestion and improves the performance of the network by employing statistical scheduling and multiple paths for different type packets. Also, queue-wise preemption technique was adopted to maximize the utilization of the links. A topology of ZigBee network was simulated using NS2, and it revealed that the proposed SMQP scheme significantly improves the network in terms of end-to-end delay and packet delivery ratio.
In the future, we will further enhance the proposed scheme with more sophisticated load balancing scheme and queuing discipline, along with the study on other performance metrics including throughput. We will also improve the queuing model of the proposed scheme accordingly. The proposed scheme will be extended to be applied to the virtualized network environment such as software defined networking (SDN). Here the schemes for dynamic path selection and load balancing will be developed.
This work was partly supported by Institute for Information & communications Technology Promotion (IITP) grant funded by the Korea government (MSIT) (No. 2016-0-00133, Research on Edge computing via collective intelligence of hyperconnection IoT nodes), Korea, under the National Program for Excellence in SW supervised by the IITP (Institute for Information & communications Technology Promotion) (2015-0-00914), Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2016R1A6A3A11931385, Research of key technologies based on software defined wireless sensor network for realtime public safety service, 2017R1A2B2009095, Research on SDN-based WSN Supporting Real-time Stream Data Processing and Multiconnectivity), the second Brain Korea 21 PLUS project, and Samsung Electronics.
- 1.Zigbee Alliance. (2017). Retrieved January 15, 2017 from http://www.zigbee.org/.
- 3.Al-Ghamdi, B., Ayaida, M., & Fouchal, H. (2015). Scheduling approaches for wireless sensor networks. In 2015 15th international conference on innovations for community services (I4CS) (pp. 1–6). IEEE.Google Scholar
- 5.Marina, M. K., & Das, S. R. (2001). On-demand multipath distance vector routing in ad hoc networks. In Ninth international conference on network protocols (pp. 14–23). IEEE.Google Scholar
- 6.Yuan, Y., Chen, H., & Jia, M. (2005). An optimized ad-hoc on-demand multipath distance vector (AOMDV) routing protocol. In 2005 Asia-Pacific conference on communications (pp. 569–573). IEEE.Google Scholar
- 7.Zhong, D., Ji, W., Liu, Y., Han, J., & Li, S. (2011). An improved routing algorithm of Zigbee wireless sensor network for smart home system. In 2011 5th international conference on automation, robotics and applications (ICARA) (pp. 346–350). IEEE.Google Scholar
- 10.Karim, L., Nasser, N., Taleb, T., & Alqallaf, A. (2012). An efficient priority packet scheduling algorithm for wireless sensor network. In 2012 IEEE international conference on communications (ICC) (pp. 334–338), IEEE.Google Scholar
- 11.Yu, X., Xiaosong, X., & Wenyong, W. (2009). Priority-based low-power task scheduling for wireless sensor network. In 2009 international symposium on autonomous decentralized systems.Google Scholar
- 12.Edalat, N., Xiao, W., Tham, C.-K., Keikha, E., & Ong, L.-L. (2009). A price-based adaptive task allocation for wireless sensor network. In IEEE 6th international conference on mobile adhoc and sensor systems, 2009. MASS’09 (pp. 888–893). IEEE.Google Scholar
- 13.Momeni, H., Sharifi, M., & Sedighian, S. (2009). A new approach to task allocation in wireless sensor actor networks. In First international conference on computational intelligence, communication systems and networks, 2009. CICSYN’09 (pp. 73–78). IEEE.Google Scholar
- 14.Tirkawi, F., & Fischer, S. (2008). Adaptive tasks balancing in wireless sensor networks. In 3rd international conference on information and communication technologies: From theory to applications, 2008. ICTTA 2008 (pp. 1–6), IEEE.Google Scholar
- 15.Zhao, Y., Wang, Q., Wang, W., Jiang, D., & Liu, Y. (2009). Research on the priority-based soft real-time task scheduling in TinyOS. In International conference on information technology and computer science, 2009. ITCS 2009 (Vol. 1, pp. 562–565). IEEE.Google Scholar
- 16.Christinal, B., & Vinodhini, V. (2014). An efficient multitask scheduling technique for the improvement of WSN with network lifetime & delay constraint. International Journal of Advanced Research in Computer Science, 5(6), 225–230.Google Scholar
- 17.Preemption and Context Switching | The Linux Process Scheduler | InformIT. (n.d.). Retrieved January 16, 2017 from http://www.informit.com/articles/article.aspx?p=101760&seqNum=3.
- 18.Preemption (computing). (2017). Retrieved January 5, 2017 from https://en.wikipedia.org/wiki/Preemption_(computing).
- 19.Kim, K.-I., Park, S., Park, H., & Ham, Y. H. (2008). Reliable and real-time data dissemination in wireless sensor networks. In Military communications conference, 2008. MILCOM 2008 (pp. 1–5). IEEE.Google Scholar
- 20.Lu, C., Blum, B. M., Abdelzaher, T. F., Stankovic, J. A., & He, T. (2002). Rap: A real-time communication architecture for large-scale wireless sensor networks. In Eighth IEEE real-time and embedded technology and applications symposium, 2002. Proceedings (pp. 55–66). IEEE.Google Scholar
- 22.Mizanian, K., Hajisheykhi, R., Baharloo, M., & Jahangir, A. H. (2009). RACE: A real-time scheduling policy and communication architecture for large-scale wireless sensor networks. In Seventh annual communication networks and services research conference, 2009. CNSR’09 (pp. 458–460). IEEE.Google Scholar
- 23.Lin, K., Zhao, H., Yin, Z., & Bi, Y. (2007). An adaptive double ring scheduling strategy based on tinyos. Journal-Northeastern University Natural Science, 28(7), 985.Google Scholar
- 24.Lee, E.-M., Kashif, A., Lee, D.-H., Kim, I.-T., & Park, M.-S. (2010). Location based multi-queue scheduler in wireless sensor network. In 2010 the 12th international conference on advanced communication technology (ICACT) (Vol. 1, pp. 551–555). IEEE.Google Scholar
- 25.Karimi, E., & Akbari, B. (2011). Improving video delivery over wireless multimedia sensor networks based on queue priority scheduling. In 2011 7th international conference on wireless communications, networking and mobile computing (WiCOM) (pp. 1–4). IEEE.Google Scholar
- 27.Lafta, H. A., & Salman, F. M. (2014). Optimal path selection in ad hoc (MANET) by using genetic fuzzy petri net. Researcher, 6(8), 31–44.Google Scholar
- 28.Anon. (2003). Stochastic Modelling and Applied Probability. In Applied Probability and Queues. https://doi.org/10.1007/b97236.
- 30.Zheng, J., & Lee, M. J. (2006). A comprehensive performance study of IEEE 802.15. 4. In Sensor network operations (pp. 218–237). IEEE Press, Wiley Interscience.Google Scholar
- 31.NS2. (n.d.). Retrieved February 16, 2017 from http://www-ee.ccny.cuny.edu/zheng/pub.