A Survey on Congestion Control at Transport Layer in Wireless Sensor Network

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS Volume 1, Number 1, December 2014 OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS A Survey on Congestion Control at Transport Layer in Wireless Sensor Network Prakul Singhal*, Anamika Yadav Department of Electrical Engineering, NIT Raipur, C.G. (India)-492010. *Corresponding author: prakul 2008@yahoo.co.in Abstract: Congestion control in a wireless sensor network is a vital issue in the present scenario. In this paper, a comprehensive survey on congestion control schemes at transport layer in a wireless sensor network is presented. At transport layer, the protocols are based on two approaches viz. Reliability and congestion control. Various protocols in each category have been discussed. Thorough survey on various congestion control schemes has been done. Latest research work including artificial intelligence based techniques e.g. fuzzy logic as well as neural network adaptive based congestion control scheme have been included. Keywords: Wireless Sensor Network; Congestion Control Protocols; Reliability Protocols; Buffer Occupancy; Rate Adjustment 1. INTRODUCTION Wireless sensor networks (WSN) are the collection of hundreds or thousands of sensor nodes distributed randomly in a geographical area. Each sensor node has the ability to sense, communicate and compute. They have processor, memory, transceiver and sensors with limited battery life. WSN are used to monitor events in a region where humans are incapable to monitor. It cooperatively monitors different conditions, like sound, temperature, vibration, pressure and motion etc within a region. The data collected from the region are sent to the sink nodes that connect the sensor network with one or more users as shown in Figure 1. The main characteristics of a WSN are mobile nodes, tolerant node failures, scalability to large scale of deployment, nodes heterogeneity and low power consumption. It can be used in many applications such as habitat monitoring, health care security surveillance, target tracking, military application and etc. However, there are some problems that need to be overcome, for example, reliability data delivery and congestion control. These problems are controlled through the help of protocols. Some protocols provide only reliability [1 3], some provide only congestion control [2] and some others provide both reliability and congestion control [2]. 17

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS Figure 1. Wireless Sensor Networks 2. TRANSRORT PROTOCOLS There are two major functions in the transport protocol of WSNs i.e. reliability and congestion control. Reliable data delivery means packet should reach the destination in multi-hop WSN; if not then proper mechanism is applied to recover the lost packet. Congestion occurs when there is many to one type scenario or if packet service time is more as compared to packet arrival time. This creates a situation of traffic which exceeds the network capacity. Due to congestion in a WSN, congested nodes initiate to drop data packets or the delay of the packets due to large and filled queue. Dropping of packets cause wastage of energy and also affect reliability. The reliability and congestion control can be done by hop-by-hop or by end-to-end approach. 2.1 Protocol for Reliability Reliability means accurate delivery of data packet from source to destination. It can be either Upstream or Downstream Reliability. In Upstream reliability, data packets move from source to sink node whereas in downstream reliability it moves from sink to source node. If data packet moves in both directions then it is known as Bidirectional reliability. Reliability protocols are classified into two general classes namely (1) packet-based and (2) event-based. In packet-based reliability, lost packets are detected at sink or any intermediate nodes. And for achieving reliability, retransmission of lost packet is carried out. Whereas, in event-based approach the reliability achieved at sink node is signaled back to source, which is done by the help of end-to-end source rate adjustment. Various protocols for achieving reliability are: 18 1. ERTP: Energy-efficient Transport Protocol (ERTP) [4] is used for data streaming applications, where sensor monitoring are send through many sensor nodes to sink nodes i.e. base station.

A Survey on Congestion Control at Transport Layer in Wireless Sensor Network It is a packet-based reliability protocol which applies statistical metric aiming to deliver more number of packets to sink as compared to the defined threshold. It reduces energy dissipation by using end-to-end approach which dynamically controls the reliability at each hop. It controls the maximum number of retransmissions dynamically at each sensor node. For recovery, it uses Stop-and-Wait Hop-by-Hop Implicit Acknowledgment (iack). The sink node explicitly sends the acknowledgment signal to the source node, if the signal does not reach within a particular time then source node retransmits the unacknowledged packet. 2. PSFQ: Pump Slowly Fetch Quickly (PSFQ) [5] is a protocol used specially in downstream multicast dynamic code update. It can also be used for unicast communication. In this, data is reconstructed over each node. It is a hop by hop protocol, thus having no guaranty of end to end reliability in some scenarios. It has three functions: message transferring (pump operation), relay-initiated error recovery (fetch operation), and selective status reporting (report operation). It uses localized recovery process among immediate neighbors for achieving loose delay bound. As it is downstream protocol so it strictly manages and control reliability in reverse direction i.e. from sink to source node. Therefore, useful resources are wasted if we use it for forward direction. 3. RMST: The Reliable Multi-Segment Transport (RMST) [6] is a Selective NACK-based for directed diffusion. It provides guaranteed delivery, fragmentation and assembly to the required applications. It detects loss of packet at sink node from where a unicast message is send towards source node for notifying the missing packet. RMST divides the data transport from source to sink into Routing and Message Loss Detection (MLD). For achieving reliability, it uses hop-by-hop approach which sends a timer driven NACK for the missing packet to the previous node. It reduces end-to-end retransmission by storing unacknowledged packet in their caches. Automatic Repeat request (ARQ) is used to retransmit lost packet. The receivers detect the lost packet and send NACK for recovering it. 4. RBC: Reliable Bursty Convergecast (RBC) [7] forwards the data packet continuously using a window-less block acknowledgement scheme which copies the acknowledgement. It uses a large memory of sensors, so sensor nodes require large memory for using RBC protocol. In RBC, hop-by-hop approach with differentiated contention control is used. This ensures that packets will be retransmitted a few number of times. This works well in single event converge cast whereas in continuous event converge cast, it can generate fresh packet continuously till event repeats itself. 5. Garuda: GARUDA [8] is a reliable downstream protocol which achieves reliability from sink to source. It uses a two-stage NACK recovery process by dividing each node as a core member or non-core members. Wait-for-First-Packet (WFP) pulse is used for creating first packet delivery which generates core infrastructure. The first packet delivery calculates the number of hop from the sink to the particular node. The nodes which come under the path of hop count become the core member. By the election process, core nodes are created within the network, which should connect with at least one upstream core node. For overcoming under utilization scenario, it uses out-of-order strategy. During non-core recovery phase, non-core nodes request retransmission from the core nodes. After that, it listens to all retransmission from its core node and sends their retransmission request. 6. DTC: Distributed TCP Caching (DTC) [9] is an enhanced version of TCP. It compresses the header and provides caches at few selected intermediate nodes which help in improving transmission capacity. It leaves end point of communication unchanged as it is fully compatible with TCP. It uses the AIMD algorithm for adjusting transmission window and hop-by-hop loss recovery scheme 19

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS Table 1. Summary of reliable protocols Protocols Reliability Direction Reliability level Recovery approach Notification ERTP [4] Upstream Packet based Hop-by-hop ACK, iack PSFQ [5] Downstream Packet based Hop-by-hop NACK RMST [6] Upstream Packet based Hop-by-hop, End-to-end NACK RBC [7] Upstream Packet based Hop-by-hop iack, NACK GARUDA [8] Downstream Packet and destination based Two stage loss recovery NACK DTC [9] Upstream Packet based Hop-by-hop ACK DTSN [10] Upstream Packet based End-to-end ACK, NACK IPSFQ [11] Downstream Packet based Hop-by-hop NACK EEHRTP [12] Upstream Packet based End-to-end Hierarchical iack [13] Upstream Packet based Hop-by-hop iack for lost packets which can be received from intermediate node caches. Those segments which are not acknowledged at link layer, are saved into the cache of intermediate node in next hop, and retransmitted when transmission time is out. 7. DTSN: Distributed Transport for Sensor Network (DTSN) [10] is an energy-efficient hop-by-hop reliable transport protocol supporting both full and differentiated reliability and employs selective repeat ARQ to improve energy efficiency. In order to reduce the overhead of packet in DTSN, the loss recovery process is controlled by the source node. It sends Explicit Acknowledgment Request (EAR) from receiver to sender which can be either ACK or NACK according to retransmission interval. The receiver node finds the lost packet by the help of missing sequence number and makes a list of lost packets keeping it until source node sends an EAR to it. 8. IPSFQ: Improved Pump Slowly Fetch Quickly (IPSFQ) [11] is an improved version of PSFQ [5] protocol. This protocol removes the short coming of PSFQ that are in-sequence forwarding of data packets, pump and repairing operations. By improving these short coming IPSFQ performs better in terms of error tolerance and average latency. 9. EEHRTP: Energy Efficient Hierarchical Reliable Transport Protocol (EEHRTP) [12] increases the network lifetime by controlling redundant data transmission. It achieves end-to-end reliability using hierarchical implicit acknowledgment. It uses Stop-and-Wait Hop-by-Hop implicitly acknowledgement for loss recovery and timeout scheme for packet retransmission. The packet loss rate and packet delivery rate show that it is better than ERTP [4] in terms of energy conserving. 10. In [13] author proposed a reliable data transfer protocol which uses cross layer optimization. This protocol is an energy-efficient hop-by-hop reliable protocol, which retransmit the packet if timeout happen before reaching to sink node. Hop-by-hop reliability is done by using four state conditions. The summary of various reliable protocols stated above is given in Table 1 below. 2.2 Protocols for congestion control 20 Congestion occurs in WSN due to many-to-one nature of traffic. When such an event occurs, many sensor nodes sense it and send packets toward one or more sinks. Congestion may also occur due to limited wireless bandwidth of sensor network. It can be either node-level congestion or link-level congestion as shown in Figure 2. Both of them lead to wastage of the energy of the nodes. Congestion occurring

A Survey on Congestion Control at Transport Layer in Wireless Sensor Network Figure 2. Node level and link level congestion in WSN is different from that occurring in the wired networks. That s why existing protocols such as TCP are not capable to handle congestion in WSN. The congestion pattern is different for upstream, downstream and bidirectional data flow. Congestion causes overall degradation of channel quality as well as packet delivery ratio, it also causes buffer to overflow and increased delays which lead to reduction in throughput. Thus, congestion control is vital for avoiding congestion in the network, which can improve network performance. Generally, congestion control algorithms in WSNs employ two techniques for controlling and avoiding the congestion viz. traffic control and resource control. The traffic control protocols use traffic control method to adjust the rate of the source node for controlling congestion in the network. On the other hand, resource control algorithms employ redundant nodes, which are not in the initial path from source to sink, so that data can flow properly within the network. Congestion control mechanisms [14] are composed of three components: congestion detection, congestion notification and reporting rate adjustments. 2.2.1 Congestion detection Congestion detection is the process of detecting the presence as well as the location of congestion in the network. Different parameters are considered for detecting the congestion by different protocols. Some protocols use only Buffer Occupancy; some others use Packet Rate or Packet Service Time versus Packet Inter-Arrival Time and some protocols consider the combination of above parameters. 2.2.2 Congestion Notification In congestion notification, congestion information is passed by congested node either to the intermediate node or to the source node or to sink node or to all other nodes after detecting congestion. The congestion information can be sent explicitly or implicitly according to the protocol. Some protocols notify the congestion by setting Congestion Notification (CN) bit in packet header. 21

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS 2.2.3 Congestion Avoidance Congestion can be avoided by simply stopping sending packets into the network, by reducing its sending rate, or if all the intermediate nodes decrease their sending flow rate to their immediate neighbours. Congestion avoidance uses three different types of techniques: Rate Adjustment, Traffic Redirection and Polite Gossip Policy. Different protocols for congestion control are given in the following subsections: 1. CODA: Congestion Detection and Avoidance (CODA) [15] detects congestion by monitoring buffer occupancy as well as wireless channel load. It is both hop-by-hop and end-to-end according to the situation. The packets are dropped by the preceding node of congested node for controlling the congestion. It works similarly to TCP during packet loss, adjusting the traffic by using additive increase multiplicative decrease (AIMD) technique. CODA designs both open and closed-loop rate adjustment. Congestion is notified by the help of broadcasting the message to all nodes. In dynamic scenarios, fairness is achieved by using dynamic weight adaptation algorithm. 2. FUSION: Fusion [16] detects congestion by considering the queue length. The congested node sets a CN (congestion notification) bit in the header of each outgoing packet. After CN bit is set, the entire neighbouring node stop sending packet to the congested node which helps in clearing the queue packets in the buffer of congested node. It uses prioritized MAC algorithm and traffic rate adjustment of source nodes. It has been found that it gets good throughput as well as fairness even at high load. 3. CCF: Congestion Control and Fairness (CCF) routing scheme [17] detects congestion by finding packet service time from which it calculates current service rate. If service rate is more as compared to arrival rate of each intermediate node then there is congestion in that intermediate sensor node. It controls congestion by using hop-by-hop approach which calculates exact rate adjustment by the help of current service time and number of child nodes. As the rate adjustment depends on packet service time only, it may lead to low utilization when some sensor nodes have limited traffic or there is a significant packet error rate. 4. PCCP: Priority-based congestion control protocol (PCCP) [18, 19] detects congestion level by the help of congestion degree and node priority index, in which congestion degree is calculated by using packet inter arrival time along with packet service time. It applies hop-by-hop approach with cross layer optimization. It reduces the buffer occupancy which helps in reducing packet loss and improves energy-efficiency. It also gets high link utilization with low packet delay. 5. SENTCP: Congestion is detected by SENTCP [20] using average local packet service time and inter arrival time for calculating congestion degree. It is an open-loop hop-by-hop congestion control protocol. In SENTCP, a feedback containing local congestion degree with buffer length is send to previous node by each intermediate sensor node to adjust traffic rate. 6. ART: ART [21] controls congestion by selecting subset of sensor nodes known as essential node which covers whole area of network. These essential nodes sense the network state in energy efficient way. It is end-to-end type upstream congestion control protocol in which only essential nodes are considered for the transfer of reliable information in upstream and downstream. 22 7. TRICKLE: Trickle [22] uses Polite Gossip technique for controlling traffic. In this protocol, a summary of data is broadcasted by each node after regular interval of time and politely suppress its own data information. Information is suppressed if same information is received more number of

A Survey on Congestion Control at Transport Layer in Wireless Sensor Network times as compared to threshold by the neighboring nodes. On the other hand, if nodes hear some new information than it broadcasts this information repeatedly with shorter interval of time, so that new information is broadcasted in the system quickly. 8. SIPHON: Siphon [23] also infers congestion by the help of queue length of intermediate nodes, as well as traffic redirection but it uses traffic redirection in place of rate adjustment technique for handling congestion. It uses the same mechanism of CODA with additional features for controlling the congestion in the secondary network having virtual sinks only. Virtual station is used to send message to all nodes for indicating that sending rate is high indicating congestion. A message containing signature byte is broadcasted by virtual sink to the secondary network on receiving some notification from base station. This message notifies the secondary network that something is not correct. 9. RCRT: Rate controlled reliable transport protocol (RCRT) [24] detects congestion by the help of base station which explicitly sends request to sensor nodes for missing packets. It provides end-to-end reliability with NACK based loss recovery scheme. The base station notifies the nodes when system enters in congestion state and repairing time is greater than a round-trip time. Base station itself decides how much traffic rate is adjusted to control the congestion by using AIMD scheme. 10. STCP: Sensor Transmission Control Protocol (STCP) [25] is a transport layer protocol in which most of the functionality is implemented at the sink node or base station. It is reliable and scalable protocol which uses multi-purpose sensor nodes but it does not provide any explicit mechanism for congestion control as well as in terms of delay ACK/NACK, it may not be feasible for reliability purpose. 11. IDCCP: Improved Datagram Congestion Control Protocol IDCCP [26] is used to control congestion in wireless multimedia network. It uses Congestion Control Identifier (CCID) for controlling purpose. An optional ACK is used in it for achieving reliability in the system. Different rate adjustment levels are used in different state. During normal State and congestion state, if there is no packet loss, the rate is adjusted according to basertt/avgrtt. If there is packet loss during congestion state, then traffic rate is halved. When system is in error state, rate is adjusted according to the variable β ranging from 0.5 to 1. In Failure State, a special packet known as probe is send for monitoring the system. 12. MCCP: Multi-event congestion control protocol (MCCP) [27] detects congestion by the help of two parameters: buffer size and packet delivery time between two sensor nodes. It uses a TDMA schedule for assigning the rate to the nodes. For scheduling purpose, it requires time synchronization, which increases complexity and overhead in the network. 13. CONSEQ: Control of Sensor Queues (CONSEQ) [28] uses a metrics for estimating congestion in the network. It uses load balancing technique during low congestion. Congestion is controlled by each node locally in their neighbourhood. This reduces congestion which improves delay as well as energy consumption in the network. It uses fuzzy control theory for dynamically rate adaption using effective queue length. As it detects congestion by the help of locally detecting technique, it responds to congestion quickly with lesser overhead. 14. FCCTF: Fairness Congestion Control for a distrustful WSN using Fuzzy logic (FCCTF) [29] control congestion by isolating malicious node. It uses dynamic threshold trust value (TTV) to block or unblock the malicious node using fuzzy logic. It performs differently in different circumstances 23

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS according to number of lost packets. When number of packets is lost, it increases TTV value so that more number of nodes is detected as malicious node. Malicious nodes do not take part in the network. When there is no packet lost then it reduces the TTV value as well as releases some malicious nodes so that they can take part in the network. As number of nodes increases, more packets start flowing within the network. 15. FBACC: The Fuzzy Based Adaptive Congestion Control (FBACC) [30] is proposed which uses buffer occupancy, participants and traffic rate. FBACC provides a fuzzy logic based congestion estimation, smart way to drop packet in case of congestion up to acceptable quality level and most importantly, it regulates traffic rate. It improves performance of network over other protocols which are presently available. Performance of FBACC is evaluated and compared with existing schemes e.g. Event-to-Sink Reliable Transport (ESRT) and Fuzzy Logic Based Congestion Estimation (FLCE) in terms of congestion detection, packet loss and energy. FBACC detects the network congestion more precisely and uniformly and also adapts to current traffic rate with respect to product of previous participants with previous traffic rate to reduce the packet loss. Finally, FBACC saves energy of retransmission because packet loss is minimal due to traffic adaption. 16. PHSA: Probability based Hop Selection Approach (PHSA) [31] controls congestion by controlling the resources of the network. This protocol finds the path cost between node and each sink node by exchanging information and sending and receiving data steps. Computed cost is exchanged between nodes which help in electing the node for next hop. Thus, resources are adjusted for controlling the congestion. This protocol also gives efficient power consumption with high packet delivery ratio. 17. HTAP: Hierarchical Tree Alternative Path Algorithm (HTAP) [32] controls congestion by dynamically switching to alternate path on the basis of local information. Local information contains congestion level of their neighbouring nodes by the help of adaptive method. Adaptive method infers the congestion level by the help of buffer occupancy with duration. In this protocol, each node is only connected to those nodes which are in upstream direction i.e. from source to sink node. 18. SUIT: Sensor fuzzy-based image transport (SUIT) [33] is a fuzzy based congestion control protocol. It sends maximum number of frame to sink node with lower quality during congestion state. This protocol uses cross layer technique to interact with different layers. It uses three indicators to detect congestion that are ratio of incoming to outgoing packets, number of active neighbour and buffer occupancy of the parent node. It applies fuzzy on the indicators and if congestion is detected than it adjust the rate to control congestion. 19. NNBCD: Neural Network Based Congestion Detection protocol (NNBCD) [34] trains the neural network by the help of three parameters i.e. buffer occupancy, number of participant and traffic rate for detecting congestion in the congested WSN. These parameters are extracted from NAM trace file by using AWK script. This protocol is able to detect congestion accurately and efficiently. Further this technique can be utilized for controlling the congestion by adapting the traffic rate. The summary of various congestion protocols stated above is given in Table 2 below. 3. FUTURE WORKS 24 The protocols discussed above have been effective in improving reliability and controlling congestion in transport layer of WSN. Some reliable protocols used hop-by-hop or end-to-end recovery approach and some of them used both approaches. Reliable protocol can be upstream or downstream. And in congestion

A Survey on Congestion Control at Transport Layer in Wireless Sensor Network Table 2. Summary of Congestion protocols Congestion Protocols Detection Notification Control technique CODA [15] Buffer occupancy, channel load Explicitly AIMD FUSION [16] Queue length Implicitly Stop and start CCF [17] Packet service time Implicitly Exact rate PCCP [18, 19] Congestion degree, packet inter arrival and service time Implicitly Exact rate SENTCP [20] average local packet service time, inter arrival time Explicitly Rate adjustment ART [21] Service time Implicitly Rate adjustment TRICKLE [22] - - Polite gossip SIPHON [23] Queue length - Traffic redirection RCRT [24] Packet loss, recovery dynamics Explicitly AIMD-like STCP [25] Queue length Implicitly AIMD-like IDCCP [26] - Explicitly Rate adjustment MCCP [27] Buffer size, packet delivery ratio - Rate adjustment CONSEQ [28] Effective and virtual queue length - Rate adjustment FCCTF [29] Buffer overflowing, forwarding rate vs receiving rate - Malicious nodes adaption FBACC [30] Participants, buffer occupancy and traffic rate Explicitly Rate adjustment PHSA [31] Queue drop, local cost Explicitly Resource adjustment HTAP [32] Buffer occupancy, duration - Alternate path SUIT [33] Number of active neighbour, buffer occupancy Rate adaption - and incoming to outgoing packet ratio and quality adaption NNBCD [34] Buffer occupancy, participants and traffic rate - - control protocols detect congestion by the help of certain parameters such as buffer occupancy, channel load, traffic rate, packet loss, participants etc. They control the congestion by using various adjustment techniques such as traffic rate adjustment, traffic redirection, polite gossip, resource adjustment etc. However the protocol discussed above can be further improved by applying various methods: 1. Computational intelligence [35] is an untouched field till now, only fuzzy application of computational intelligence is used in few paper but other technique like neural network, genetic algorithm, reinforcement learning etc. are also appropriate for improving QoS of WSN at transport layer. 2. Currently all protocol start congestion mitigation method after detecting of congestion state but a better protocol can be developed which mitigate the congestion before happing of congestion. So it can saves lot of energy and increase throughput with better resource utilization. 3. Existing protocols either adapt traffic or redirect traffic flow not both. Hybrid protocol can be developed which can use both traffic rate adaption with traffic redirection. This can be performed by using cross layer approach, as they can interact with different layer by escaping virtually strict boundaries between layers. 4. CONCLUSION In this paper, a survey of various protocols used for congestion control (CC) in Wireless Sensor Networks is presented. Different types of protocols which are commonly used for congestion detection and control in WSN and also some other protocols which have been recently proposed are discussed in detail. Fairness Congestion Control protocol for a distrustful WSN using Fuzzy logic (FCCTF) is found to be an effective protocol as it uses dynamic threshold value to control congestion. As can be seen from 25

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS Table 2, FCCTF considers both packet drop using buffer overflow and traffic rate by using forwarding rate vs receiving rate. It uses one of the Artificial Intelligence techniques: Fuzzy logic for estimating the dynamic threshold value by considering the input parameters, which helps in controlling the congestion with better utilization of channel. It performs differently in different circumstances according to number of lost packets. The Fuzzy Based Adaptive Congestion Control (FBACC) has proved to be very effective protocol for wireless multimedia sensor network, as it maintains the network in medium congestion level by using number of participant along with buffer occupancy and traffic rate. FBACC provides a fuzzy logic based congestion estimation, smart way to drop packet in case of congestion up to acceptable quality level and most importantly it regulate traffic rate. FBACC saves energy of retransmission because packet loss is minimal due to traffic adaption. Thus, it can be concluded that till date many effective protocols have been proposed to detect and control congestion in transport layer for WSN. Nevertheless, there is still scope for developing more effective protocols using artificial intelligence, cross layer technique etc. References [1] F. Yunus, N.-S. Ismail, S. H. Ariffin, A. Shahidan, N. Fisal, and S. K. Syed-Yusof, Proposed Transport Protocol for Reliable Data Transfer in Wireless Sensor Network (WSN), in Modeling, Simulation and Applied Optimization (ICMSAO), 2011 4th International Conference on, pp. 1 7, IEEE, 2011. [2] S. Sridevi and M. Usha, Taxonomy of Transport Protocols for Wireless Sensor Networks, in Recent Trends in Information Technology (ICRTIT), 2011 International Conference on, pp. 467 472, IEEE, 2011. [3] D. N. Wategaonkar and V. S. Deshpande, Characterization of Reliability in WSN, in Information and Communication Technologies (WICT), 2012 World Congress on, pp. 970 975, IEEE, 2012. [4] T. Le, W. Hu, P. Corke, and S. Jha, ERTP: Energy-efficient and Reliable Transport Protocol for data streaming in Wireless Sensor Networks, Computer Communications, vol. 32, no. 7, pp. 1154 1171, 2009. [5] C.-Y. Wan, A. T. Campbell, and L. Krishnamurthy, PSFQ: a reliable transport protocol for wireless sensor networks, in Proceedings of the 1st ACM international workshop on Wireless sensor networks and applications, pp. 1 11, ACM, 2002. [6] F. Stann and J. Heidemann, RMST: Reliable Data Transport in Sensor Networks, in Sensor Network Protocols and Applications, 2003. Proceedings of the First IEEE. 2003 IEEE International Workshop on, pp. 102 112, IEEE, 2003. [7] H. Zhang, A. Arora, Y.-r. Choi, and M. G. Gouda, Reliable bursty convergecast in wireless sensor networks, vol. 30, pp. 2560 2576, Elsevier, 2007. [8] R. Sivakumar and I. F. Akyildiz, GARUDA: Achieving Effective Reliability for Downstream Communication in Wireless Sensor Networks, IEEE Transactions on Mobile Computing, vol. 7, no. 2, pp. 214 230, 2008. [9] A. Dunkels, J. Alonso, T. Voigt, and H. Ritter, Distributed TCP Caching for Wireless Sensor Networks, SICS Research Report, 2004. [10] B. Marchi, A. Grilo, and M. Nunes, Dtsn: Distributed transport for sensor networks, pp. 165 172, 2007. [11] C. K. Rupani and T. C. Aseri, An improved transport layer protocol for wireless sensor networks, Computer Communications, vol. 34, no. 6, pp. 758 764, 2011. 26

A Survey on Congestion Control at Transport Layer in Wireless Sensor Network [12] P. Mohanty and M. R. Kabat, Energy Efficient Hierarchical Reliable Transport Protocol in Wireless Sensor Networks, Global Journal on Technology, vol. 3, pp. 61 67. [13] F. Yunus, N.-S. Ismail, S. H. Ariffin, A. Shahidan, N. Fisal, and S. K. Syed-Yusof, Proposed Transport Protocol for Reliable Data Transfer in Wireless Sensor Network (WSN), in Modeling, Simulation and Applied Optimization (ICMSAO), 2011 4th International Conference on, pp. 1 7, IEEE, 2011. [14] B. Sharma and T. Aseri, A Comparative Analysis of Reliable and Congestion-Aware Transport Layer Protocols for Wireless Sensor Networks, ISRN Sensor Networks, 2012. [15] C.-Y. Wan, S. B. Eisenman, and A. T. Campbell, CODA: congestion detection and avoidance in sensor networks, pp. 266 279, 2003. [16] L. A. Freitas, A. R. Coimbra, V. Sacramento, S. Rosseto, and F. M. Costa, A Data Fusion Protocol in Wireless Sensor Networks for Controlled Environment, in INFOCOM Workshops 2009, IEEE, pp. 1 2, IEEE, 2009. [17] C. T. Ee and R. Bajcsy, Congestion control and fairness for many-to-one routing in sensor networks, in Proceedings of the 2nd international conference on Embedded networked sensor systems, pp. 148 161, ACM, 2004. [18] C. Wang, K. Sohraby, V. Lawrence, B. Li, and Y. Hu, Priority-based Congestion Control in Wireless Sensor Networks, in Sensor Networks, Ubiquitous, and Trustworthy Computing, 2006. IEEE International Conference on, vol. 1, pp. 148 161, IEEE, 2006. [19] D. Patil and S. N. Dhage, Priority-based Congestion Control Protocol (PCCP) for controlling upstream congestion in Wireless Sensor Network, in Communication, Information & Computing Technology (ICCICT), 2012 International Conference on, pp. 1 6, IEEE, 2012. [20] C. Wang, K. Sohraby, and B. Li, SenTCP: A hop-by-hop congestion control protocol for wireless sensor networks, in IEEE INFOCOM, pp. 107 114, 2005. [21] N. Tezcan and W. Wang, ART: an asymmetric and reliable transport mechanism for wireless sensor networks, International Journal of Sensor Networks, vol. 2, no. 3-4, pp. 188 200, 2007. [22] P. A. Levis, N. Patel, D. Culler, and S. Shenker, Trickle: A self-regulating algorithm for code propagation and maintenance in wireless sensor networks. Computer Science Division, University of California, 2003. [23] C.-Y. Wan, S. B. Eisenman, A. T. Campbell, and J. Crowcroft, Siphon: Overload traffic management using multi-radio virtual sinks in sensor networks, in Proceedings of the 3rd international conference on Embedded networked sensor systems, pp. 116 129, ACM, 2005. [24] J. Paek and R. Govindan, Rcrt: Rate-controlled reliable transport for wireless sensor networks, in Proceedings of the 5th international conference on Embedded networked sensor systems, pp. 305 319, ACM, 2007. [25] Y. G. Iyer, S. Gandham, and S. Venkatesan, STCP: A Generic Transport Layer Protocol for Wireless Sensor Networks, in Computer Communications and Networks, 2005. ICCCN 2005. Proceedings. 14th International Conference on, pp. 449 454, IEEE, 2005. [26] L. Yong-Min, J. Xin-Hua, N. Xiao-Hong, and L. Wu-Yi, STCP: A Generic Transport Layer Protocol for Wireless Sensor Networks, in Proceedings of the 2009 Eigth IEEE/ACIS International Conference on Computer and Information Science, pp. 194 198, IEEE Computer Society, 2009. [27] H. Faisal B, C. Yalcin, S. Ghalib A, et al., A multievent congestion control protocol for wireless sensor networks, EURASIP Journal on Wireless Communications and Networking, vol. 2008, 2009. [28] C. Basaran, K.-D. Kang, and M. H. Suzer, Hop-by- Hop Congestion Control and Load Balancing in Wireless Sensor Networks, in Local Computer Networks (LCN), 2010 IEEE 35th Conference on, pp. 448 455, IEEE, 2010. [29] M. Zarei, A. M. Rahmani, R. Farazkish, and S. Zahirnia, FCCTF: Fairness Congestion Control for 27

OPEN TRANSACTIONS ON WIRELESS COMMUNICATIONS a distrustful wireless sensor network using Fuzzy logic, in Hybrid Intelligent Systems (HIS), 2010 10th International Conference on, pp. 1 6, IEEE, 2010. [30] S. Jaiswal and A. Yadav, Fuzzy based adaptive congestion control in wireless sensor networks, in Contemporary Computing (IC3), 2013 Sixth International Conference on, pp. 433 438, IEEE, 2013. [31] N. Farzaneh and M. H. Yaghmaee, Probability based hop selection approach for resource control in Wireless Sensor Network, in Telecommunications (IST), 2012 Sixth International Symposium on, pp. 703 708, IEEE, 2012. [32] C. Sergiou, Performance-aware congestion control in wireless sensor networks using resource control, in World of Wireless, Mobile and Multimedia Networks (WoWMoM), 2013 IEEE 14th International Symposium and Workshops on a, pp. 1 2, IEEE, 2013. [33] C. Sonmez, O. D. Incel, S. Isik, M. Y. Donmez, and C. Ersoy, Fuzzy-based congestion control for wireless multimedia sensor networks, EURASIP Journal on Wireless Communications and Networking, vol. 2014, no. 1, pp. 1 17, 2014. [34] P. Singhal and A. Yadav, Neural network based congestion detection, IEEE International Conference of Convergence and Technology (I2CT), Pune, p. 72, 2014. [35] R. V. Kulkarni, A. Forster, and G. K. Venayagamoorthy, Computational intelligence in wireless sensor networks: A survey, Communications Surveys & Tutorials, IEEE, vol. 13, no. 1, pp. 68 96, 2011. 28