An efficient Multi-Parent Hierarchical Routing Protocol for WSNs

Transcription

1 An efficient Multi-Parent Hierarchical Routing Protocol for WSNs Abstract Wireless sensor networks (WSNs) nodes are commonly designed to work with limited resources of memory, energy and processing; these constraints are aimed at reducing costs in order to make them suitable for a large number of applications: data mining, medicine, military, high-security industry and automation, among others. The routing protocol is one of the key components of WSNs and its features impact network performance significantly. We present an efficient Multi- Parent Hierarchical (MPH) routing protocol for wireless sensor networks; its main goal is to achieve reliable delivery of data in a single sink scenario while keeping low overhead, reduced latency and small energy consumption. The main features of MPH are self-configuration, hierarchical topology, persistence according to link quality and source routing from sink to nodes. Network performance simulations of the MPH routing protocol are carried out and compared with two popular protocols, AODV (Ad hoc On-demand Distance Vector) and DSR (Dynamic Source Routing). Results show that for the single sink scenario, the MPH protocol has an energy saving of 35% and 27% less overhead compared with the two other protocols. Finally, we present a real WSN built based on the MPH protocol, which works satisfactorily, providing an experimental demonstration of the capabilities of the protocol. I. INTRODUCTION Wireless Sensor Networks (WSNs) are being deployed in a wide range of applications to support new generation services such as digitalization of cities, telemetry and industry automation, among others. WSNs are based on low-cost devices (nodes) that are able to get information from their environment, process it locally, and communicate via wireless links to a central coordinator node. Besides, the coordinator node might also send control commands to the nodes [1]. WSNs may not rely on a predetermined structure and require the capacity of self-organization in order to deal with wireless impairments, mobility and nodes failures. One of the most efficient topologies in information delivery is the hierarchical topology. The hierarchy levels allow packet forwarding with the least number of hops, which causes fewer errors in delivery and lower delays on the transmission of a packet from source to destination [6]. Tree Routing is a classic form of routing that is restricted to parent-child links. This scheme eliminates the need for searching and updating paths and the overhead associated to the establishment of those paths [2]. It presents advantages for nodes such as energy saving and low memory consumption. However, when the networks are large and the nodes can connect and disconnect from the network due to link failures or communications features, it is necessary that the Tree Routing scheme slightly change. This is because it requires flexibility in assigning IP addresses to the network in order to become self-organized. ZigBee manages a Tree Routing scheme with pre-configured global IP addressing. Note that in this structure, node addresses never change [2]. We chose to develop a hierarchical protocol because these types of protocols are aimed at energy saving multi-hop communications and are able to cover wide areas [9]. Hierarchical protocols have scalability and robustness characteristics, providing energy savings in the network and distributing energy costs among network sensors. A great advantage of such protocols is that they carry information generally to one node, thus the communication with the coordinator or root node is simpler and more efficient. In this paper we propose an efficient Multi-Parent Hierarchical (MPH) Routing Protocol based on link-hierarchy relationships (parents and children); it establishes operational techniques aimed at energy saving and reliable delivery of information. In the scenario of interest, all data is gathered by a powerful coordinator node that can process more information, perform more complex tasks and manage routing tables that reflect the entire network topology. Based on the proposed MPH protocol, a network acquires a hierarchical topology since it represents an efficient way to route the information through the network with a minimum number of hops from source to destination. Two highly studied and widely used protocols for WSNs are AODV (Ad hoc On-demand Distance Vector) [15] and DSR (Dynamic Source Routing) [16]. They are reactive protocols that establish routes on demand and send information to the destination by the route of least number of hops that each node has at a given time in its routing table. These protocols use two procedures to find and transport traffic packets: route discovery and route maintenance. Their performance and efficiency have been compared by means of certain network metrics [12]. In our work we compare efficiency metrics observed in AODV and DSR with those of our proposed protocol, MPH. The rest of this article is organized as follow: in Section II, we describe related work. Section III, proposes an efficient Multi-Parent Hierarchical (MPH) Routing Protocol. Section IV, explains the proposed metrics and the simulation environment. Section V, shows and discuss the results. Section VI describes the network implementation. Finally, conclusions are given in section VII. mds January 12, 2014

2 II. RELATED WORK The routing protocol is an important factor on sensor networks which greatly influences network performance and node interactions. This will directly affect the network energy consumption, delay and reliability during the process of information delivery among nodes. Two commonly studied parameters in WSNs are energy and delay [7]. The the end-to-end delay is an important metric that depends on media access techniques, collisions and retransmissions. To tackle the energy-saving problem, sensor networks must be able to implement energy conservation mechanisms, enabling nodes to work collaboratively. The focus of the work cited in [11] is to describe schemes of data fusion or aggregation which represent balancing and energy consumption reduction strategies. Cuomo et al. [8] analyze the formation network process and the topology with a focus on energy consumption. They vary factors such as the number of sink nodes and their distribution along the network. They establish relationships between the number of sink nodes and the energy consumption and study static and mobile scenarios. In [17] they deal with the problem of balancing costs with a clustered spanning tree algorithm and an energy model based on transmission energy. In addition, they assign weights to the links so that the algorithm provides the best routes of the hierarchical tree. Sending packets to one or more nodes is the main job in sensor networks. Routing protocols use various schemes to find the best routes in the shortest time possible. The technique cited in [3] analyzes metrics such as packet loss, delay and energy as measures of the quality of service in order to improve overall performance in the network. They explain a traffic control scheme that provides a proper topology management. Regarding the problem of the study of static scenarios with connections and disconnections of nodes, Mavromoustakis et al., in [4], describe a scheme that balances the neighboring information network while adapting to new conditions. The unbalance links problem is treated in tree topology protocols that aim to minimize the problem of packet delay and ensure the least number of hops to the root node. A problem that arises from this point is that one of load unbalance on the links, which is reflected in an unbalanced energy depletion over the nodes. Reference [14] shows a method to balance network traffic and ensure a uniform use of the routes to the destination node. Nezhad et al. [10] propose a protocol where the collector node has a global view of the network topology. In order to maximize the network lifetime, they use a load balancing algorithm to choose the best routes. In [5] a multipath routing which uses a QoS technique on adhoc systems is proposed in order to provide load balancing, fault tolerance, and better performance of the network with protocols such as AODV. In order to compare sensor networks performance some authors have been working extensively with two important protocols in this type of networks: AODV and DSR. Under Requirement Scenario Sensor deployment Communication among nodes Radio communication Communication technique Coordinator node Application Sampling rate Flexibility Medium Environments TABLE I PRESCRIBED REQUIREMENTS. Static Medium-scale networks Many-to-one Wireless Multi-hop Robust computer with high memory capacity Remote monitoring 3-4 samples per minute Hardware and software modularity Non-line of sight propagation Indoors and outdoors theses protocols, the nodes find routes to a destination node and they can have multiple routes to the same destination. III. DESIGN OF EFFICIENT MULTI-PARENT HIERARCHICAL (MPH) ROUTING PROTOCOL The scenario to be considered here is a single sink scenario. It consists of static nodes that can send and receive information to and from the coordinator node. The coordinator node collects all information for the nodes. All nodes except the coordinator have limited resources; i.e. energy, memory, processing. A hierarchical logic topology fits well in this scenario with the coordinator holding the highest hierarchy; routing packets from nodes to coordinator can be done easily by establishing routes where each next-hop belongs to an immediate upper level in the topology until the coordinator is reached. The nodes have neighbor tables that avoid high processing costs, and routes that can be achieved with these tables ensure the least hop number to the coordinator. Table I presents the prescribed requirements in order to clarify the conditions for simulation and deployment needs to compare the performance protocols in WSNs. For the considered scenario we present design criteria for the mechanisms of the protocol that lead to the response to the requirements listed below. We assume nodes with limited processing and capacity because they have constrained memory and energy resources. When a node is disconnected or turned off, the network must have the ability to reconfigure itself rapidly in order to avoid losing information. Therefore, we propose the MPH routing protocol that creates a logical hierarchical network topology where the hierarchy of each node is given by its location level, the lower the level the lower the hierarchy held; thus, the root (coordinator node) will have the highest hierarchy. The topology is formed with routes that minimize the number of hops from a node to the coordinator; this restriction will reduce overall energy consumption and decrease delay. MPH allows a child node to have one or more parents at the next higher level. As a result, a node can share both children and parents with another node belonging to the same hierarchical level or generation. Figure 1 shows an example of a topology in which we describe the formation of the hierarchy levels of nodes in the MPH routing protocol. The formation of the hierarchy levels is as follows. Node 0 is the coordinator node. This

3 Fig. 1. Neighbor discovery and routing example. TABLE II NEIGHBOR TABLE EXAMPLE OF NODE 2. Neighbor ID Neighbor hierarchy Relationship Persistence 0 50 child none none parent parent parent 2 node represents the highest hierarchy level. The nodes within its coverage range are its children and have one hierarchy level below the coordinator node, i.e., nodes 1, 2 and 3 in this example. Subsequently, these nodes find other nodes in its coverage range that have not been assigned a hierarchy. Thus, node 1 has a child, node 4. Node 2 has three children nodes: 4, 5 and 6 nodes, because node 1 and 3 nodes are in the same hierarchy level. Node 3 has two children nodes: 6 and 7. Finally, node 5 has a child which is node 8. In this way the hierarchies are formed, the relationships parent-child are established and from here, the neighbor tables are configured. In Table II, we present an example of the neighbor table for node 2. We are going to take the same example to show how to do the routing of a packet from a node to the coordinator. We route a packet from node 6 to the coordinator node. The routing is done in ascending, through parent nodes. Node 6 has two parents: nodes 2 and 3. Node 6 can choose which route to take. For example, node 6 sends the packet to node 2. Finally, node 2 has a single parent which is the coordinator node (destination) and is where the packet will be delivered. The coordinator might also need to send packets to nodes in the network, for instance in certain applications two-way communications are required for sending control commands. In the proposed MPH protocol, we adopt a source-routing approach for traffic to be send from the coordinator to a given node because the coordinator node has more resources than the rest of the nodes. A. MPH mechanisms The authors have already defined a scenario, so we will propose a number of mechanisms that represent an efficient solution to the requirements such as self-configuration, rapid information delivery, low overhead and small energy consumption. Hierarchical logic construction is performed by routing protocol rules that are local to each node, so the network is configured depending on the connectivity of nodes in a certain time. Although the nodes are fixed, connectivity presents dynamic features due to changing conditions of the radio channel, interference or physical damage of nodes, provoking intermittent connections and disconnections. The autoconfiguration capability will allow the network to recover from the aforementioned events, offering a robust and uninterrupted service and maintains the neighbor tables updated. 1) Neighbor discovery mechanism: The neighbor table maintaining process is originated by a N D packet transmission. The sending and receiving nodes can update their neighbor tables with the respective addresses once each node receives the corresponding acknowledgment. N D packets are delivered at regular specified intervals, so that nodes rediscover their neighbors at this intervals. The response to a N D packet is called N DR (Neighbor Discovery Response) and the response to a N DR packet is another packet called NDRACK. Both, the NDR and NDRACK packets contain node hierarchy information. 2) Node hierarchy update mechanism: Each time a node is involved in a neighbor table maintenance process, it should be checked if it maintains its hierarchy or if it is going to change it. The rule is that the node should have a hierarchy one unit lower than his parent node(s), therefore, the node will select as a parent a neighbor node having the highest hierarchy among the registered nodes in the neighbor table. It can choose more than one parent if several neighbor nodes satisfy this condition. Moreover, if a node has no neighbors or all neighbors have zero hierarchy it will have zero hierarchy too. This situation can occur in the network initialization transitional stage when there are nodes that still have zero hierarchy. 3) Reactivity N D packet mechanism: This mechanism is valuable in dynamic networks, where this dynamism generates topology changes. A node changes its hierarchy as a result of updating its neighbors, before a period (regular specified interval) is completed. Then, the process to discover neighbors, sending a ND packet, is done when the topology changes. The purpose of this mechanism is to reduce the time in which nodes update their hierarchies, especially when extreme conditions occur due to loss of critical links in the logical topology. This will diminish auto-configuration network delays and reduce the packet loss problem by disconnecting links. 4) Persistent nodes mechanism: Some nodes do not respond to every N D packet. This can be due to packet collisions or low quality links, among other causes. To prevent unreliable changes, caused by node faults, in the neighbor tables we use a scheme that employs a slow erasure of a neighbor node and uses a variable called neighbor persistence. Prior to the transmission of a N D packet, persistence is

4 reduced by 1 for every registered node in the neighbor table. If a NDR packet is received from a neighbor, persistence takes a maximum value p. Therefore, if after successive p ND packet emissions the origin node did not receive any NDR packets from the node that initially had persistence p, then the route to this node is erased from the table. This mechanism solves the problem of inconstant changes in the neighbor tables. This problem can be caused by the momentary disconnection of the nodes. For example, occasionally, due to interference or packet collisions, the nodes do not respond in time to its neighbors, which then modify the neighbor tables. Thus, the node persistence mechanism gives more opportunities for weak links to keep association with the routing tables and avoids abrupt changes. 5) Source routing mechanism: This mechanism is used to route packets from the coordinator node to any node in the network. The coordinator node builds the whole topology of the network after it gathers the neighbor tables from the nodes. Then, the coordinator node is able to specify the complete path to be followed by the packet to reach its destination. Using the example in Figure 1, we will explain how to route a packet from the coordinator node to node 8. Because of the coordinator node builds the whole network topology and knows the exact paths to each of the nodes, it has the route to node 8: This route lists the nodes the packet must go through to reach its destination. B. Example It is important to note that the network link configuration depends on the implemented protocol. Figure 2 shows the possible links established when connecting nodes on AODV, DSR and MPH. In AODV the routing tables have the next hop for all the routes, while in DSR, the route is included in the packet. Thus, nodes use this information to determine the next hop in order to reach the coordinator. In MPH the routing is from children nodes to parents. The node chooses one parent (if it has more than one) and sends the packet. MPH shall guarantee always the shortest route to the coordinator node. In AODV and DSR the shortest route from the routing table is guaranteed, but this route is not necessarily the shortest route to reach the destination. We present a specific example using Figure 3 for a node that is sending a packet to the coordinator. If node 9 wants to send information to node 0, the possible routes the packet could go according to each routing protocol would be as follows. For AODV and DSR protocols the formed routes are very similar, however AODV has lifetime for those routes and the storing of these routes will be different for each protocol. Then, the possible routes that the packet would make according to the two protocols would be: , , , , , , , , , , , and so on. While the only route the packet would follow under MPH protocol is: The AODV and DSR routes are chosen by node 9 taking into account nodes save routes in all directions, while nodes in the MPH protocol save ascending routes to the coordinator (only through their parents). This makes MPH ensures the shortest route in all cases, however, AODV and DSR can be updated then a longer route. Fig. 2. Logical links for AODV, DSR and MPH. IV. PROPOSED METRICS AND SIMULATION ENVIRONMENT The authors compare MPH protocol performance with commonly used protocols in sensor networks, such as AODV and DSR. Although AODV and DSR are reactive protocols and MPH has a hierarchical configuration, we can analyze their performance with specific metrics and compare them. We use the following metrics in the comparison: overall energy consumption, delay, hop count and delivered packets. The delay and power consumption in a network are directly related to the complexity of the routing algorithms. When a routing protocol employs many algorithms and processes to send a packet, sensors will have longer delays and high energy consumption. We propose the MPH protocol with features such as hierarchical topology, source routing from sink to nodes and rapid topology reconfiguration when faced with an unexpected change. Based on the packet delivery ratio we can have an indication of the efficiency of the routing protocol. This is influenced by the number of hops that packets go through in their way to their destination; the smaller this number, the lower the packet loss will be. Route reliability is an important aspect when it is necessary that the routing protocol have the shortest routes available. A routing protocol provides reliability when nodes have on their tables (routing tables or neighbor tables) valid paths, i.e., routes that have not expired or have not been invalidated by the disconnection of a node. The overhead of the routing protocol will impact channel occupancy. The lower the overhead, the lower the probability of collisions, which will in turn decrease packet retransmissions. This ensures that information will be delivered more quickly and reliably. We built an event-driven network simulator, programmed in C++ with the ability to manage the nodes as separate

5 elements together. The nodes communicate via control and traffic packets and develop all tasks to be part of a network: starting, transmission, switching, receiving, channel listening, shutdown and consumption of the microcontroller. The aim of this work is to analyze important metrics in a WSN, as mentioned above, in connection with the proposed MPH protocol. For the simulations we consider a grid arrangement with 49 nodes. The horizontal and vertical grid separation is 5 m, and nodes have a coverage range of 8 m. We take a data rate of 250 kbps and a packet length of 22 Bytes. The load traffic for each node is 20 packets/sec. The coordinator node is located in one corner, establishing crowns or rings of nodes around the root node, as shown in Figure 3. All network nodes must forward packets to the coordinator node directly (one hop) or indirectly (multiple hops forming a route with intermediate nodes). Fig. 3. Network Topology. We assume a packet loss for each link of 1% [13]. The MAC Layer is constructed based on the unslotted Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocol [13], established by the IEEE /Zigbee standard, with the collision avoidance model and retransmissions. The Network Layer was put to the test with the different implementation of routing protocols such as AODV, DSR and our own MPH protocol. A. Results and Discussion From the discussion above, we consider the following five important metrics that are indicative of network performance: delay, energy, delivered packets, overhead and availability of routes. 1) Delay: One of the fundamental parameters for optimizing a sensor network is the end-to-end delay. The time a packet takes to reach its destination is variable due to several factors, for instance: the transmission speed, the packet size and, the delay of the packet in each hop in the route. Collisions and packet retransmissions also increase the end-to-end delay. The delay is related to the network complexity. The MPH protocol, through the election of a hierarchical topology, produces a reduction of the delays in the information process delivery. TABLE IV ENERGY MODEL. Voltage (mv) Current (ma) Time (ms) Start-up mode MCU running on 32-MHz clock CSMA/CA algorithm Switch from RX to TX Switch from TX to RX Radio in RX mode (processing and waiting) Radio in TX mode Shut down mode It is important to consider the delay involved in reorganizing the network due to changes in connectivity, for example due to new nodes or nodes that switch off or are faulty. Table III shows relevant delays obtained in the simulations of the MPH protocol compared with the AODV and DSR protocols. The first row shows the time required to complete the process of table maintenance performed by a neighbor node. With this process, each node builds its neighbor table. The second row describes the time it takes a packet to travel from the farthest node to the coordinator (these are the nodes of the last ring shown in Figure 4). The time denoted by ( ) is the time it takes AODV and DSR in route discovery process. In the third row we obtained the time it takes a packet from the nearest ring in Figure 4 (one hop) to get to its destination, for each of the protocols. Finally, in the fourth row we randomly turned off 20% of the network nodes. Again, we observed how long it takes a packet to reach the coordinator node from the farthest node in the topology. Here, the time denoted by ( ) defines the route discovery time for AODV and DSR, and the transient time for topology arrangement for MPH. This metric takes into account the self-configuring time of the network due to the dynamics of the wireless scenario, such as node disconnections. This is where we see the ability of each routing protocol to overcome topology changes and reorganize the network, reflected as the time it takes to recover from such events. 2) Energy: Energy is a valuable metric in WSN performance because here the sensors can optimize processing tasks and they should remain functional as long as possible without having to replace the batteries. The energy model implemented for the three protocols studied is presented in Table II. When MPH is used, nodes do not store routing tables, and routing is done via the optimal route. Therefore, this protocol provides large energy savings. This can be seen in Figure 4.a, where we also observe that AODV and DSR use more total energy than MPH because they handle more routing overhead causing more collisions and retransmissions. 3) Packet delivery ratio: Another interesting metric of the network performance is how many packets reach their destination successfully. We took a radius of 10m and analyzed the percentage of delivered packets for the three protocols. The value that this metric takes is a consequence of the ability of a routing protocol to reorganize the network. Besides, if the number of hops the packets pass through is smaller, there will be fewer errors in the information delivery. Results are depicted in Figure 4.c.

6 TABLE III DELAY PARAMETER AODV DSR MPH Neighbor discovery process for an average of 8 neighbors ms ms ms Average time it takes a traffic packet to reach its destination without shutdown ms and ms and ms nodes in the network (ring farthest to destination) ms ms Average time it takes a traffic packet to reach its destination without shutdown nodes in the network (ring closest to destination) ms ms ms Average time it takes a traffic packet to reach its destination with 20% shutdown ms and 84.17ms and ms ms and nodes in the network (ring farthest to destination) ms ms 4) Overhead: Reactive protocols such as AODV and DSR have low overhead because routes are discovered only when they are needed. However, MPH uses fewer control packets than AODV and DSR, thus nodes have low processing and simple management of neighbor tables. Therefore, MPH maintains neighbor tables with less control packets. This behavior can be seen in Figure 4.b. Fig. 5. % Nodes off vs % Reliable routes. TABLE V % VALID ROUTES. Fig. 4. (a) Total energy, (b) Overhead, (c) Packet delivery ratio. 5) Availability of routes: Reliable routes or valid routes are the routes that are active and the node can use for sending traffic. These routes may expire (according to the routing protocol) or may disappear from the tables due to disconnections of neighbor nodes. The most reliable routes will ensure more reliable delivery of information. We turned off a percentage of network nodes to observe nodes disconnections. In this way we could see how some routes become invalid and how nodes respond to reconfigure valid routes in the network, depending on the protocol. The % of reliable routes is presented in Figure 5. Thanks to the persistence mechanism in MPH, neighbor tables become safer as well as more reliable thanks to the DV packet reactivity mechanism, compared with AODV and DSR. This is because in AODV routes have timers that expire after certain period. On the other hand, DSR is aware that one route is obsolete only when it receives a route error message. In Table V we took a sampling period of 100 seconds. We made tests every 10 seconds in which we compute the average number of valid routes available in case the node has to send a traffic packet right at this moment. In the AODV and DSR TIME % Valid Routes AODV DSR MPH cases, occasionally some routes have just expired or some node in the route has been disconnected: these cases will result in invalid routes. On the other hand, in MPH almost all the time the neighbor tables have valid routes ready to be used. With regard to diversity of routes and hop count, we present Table VI. The MPH protocol gets routes with minimum number of hops but it has fewer routes. The advantage of MPH compared to the AODV and DSR protocols is that MPH does not require routing tables, the decision of a node to route a packet to the coordinator is very simple because a node chooses the most widely used route. On the other hand, we see that AODV and DSR have more routes to the coordinator but they do not always guarantee the shortest route. It is true that the node chooses the shortest route from its routing table but maybe this route is not the shortest to the destination. This effect can be caused due to the packet loss probability.

7 TABLE VI VALID ROUTES AND HOPS. TIME AODV DSR MPH Routes Hops Routes Hops Routes Hops V. NETWORK IMPLEMENTATION OF THE MPH PROTOCOL The MPH protocol was successfully deployed in a real network under a variety of conditions of temperature, humidity and obstacles. The number of nodes forming the network was 90 and 1 coordinator node, the farthest node to the coordinator node is 10 hops, and the average hop delay is 15 milliseconds. The area coverage was 0.16 km 2. We built the MPH protocol on top of the MAC layer of the standard using the CC2530 chip from Texas Instruments. The power transmission for each node was 4.5 dbm. Figure 6 shows one of the nodes running the MPH protocol Fig. 7. Real network. number of hops required for each node to the destination node (coordinator). Here we observe, via a histogram, the average number of hops that shows the simulator has a very similar distribution to the results of the real network. This shows that the simulator accurately reflects the network conditions. Figure 9 shows the delay as an another important metric to monitor network processes under protocol MPH. The delay allows to determine more specifically the proximity of actual results with respect to the simulated results. Both graphs show very similar distributions, showing the fidelity of the simulation results with respect to a real network in a specific environment. Fig. 6. Wireless node. Figure 7 shows the real network made with nodes built by us in order to test the MPH protocol. This same topology was used in the simulation to reproduce the results and compare them. In order to evaluate the protocol performance from both the real network and the simulator, we analyze the number of hops and delay, provided these metrics directly impact the energy model. Figure 8 shows the distribution of hops in all nodes in the actual network. These hops are taken from each node to the coordinator. Subsequently, the same network under the same conditions was implemented in the simulator in order to compare the behavior of the MPH protocol in a real environment and a simulation. We analyze the average Fig. 8. a) Hop distribution in the real network, b) Hop distribution in the simulation. VI. CONCLUSION In MPH, the coordinator node can be aware of approximately the whole topology due to the source routing mechanism. In AODV and DSR, the routes from the coordinator to some node are calculated in the same way that the other routes. So MPH has the facility that the coordinator node can access any node to send information, statistics or measurements requests. While in AODV and DSR, the coordinator has to discover the route to a specific node if it does not have it.

8 Fig. 9. a) Delay distribution in the real network, b) Delay distribution in the simulation. MPH protocol has less control packets, therefore less overhead, resulting in fewer collisions, so there will be less packet retransmissions compared with AODV and DSR. Moreover, this is reflected in the energy saving metric from MAC layer. Results for MPH protocol are encouraging because this protocol has good performance in terms of processing, fast and efficient information delivery and energy conservation. Protocols such as AODV and DSR are very efficient in terms of backup routes and connectivity from any node to any node in the network. The combination of a hierarchical topology with self-configuration and maintenance mechanisms of the MPH protocol makes the nodes optimize network processes, reduce delays, take short routes to the destination and decrease network overhead. All this is reflected in the successful delivery of information. [10] A. Nezhad, A. Miri, D. Makrakis, D., Location privacy and anonymity preserving routing for wireless sensor networks, Computer Networks 52, pp , Dec [11] H. Luo, Y. Liu, S. Das, Routing Correlated Data in Wireless Sensor Networks: A Survey, IEEE Network 21, pp , Dec [12] N. Kulkarni, R. Prasad, H. Cornean, N. Gupta, Performance Evaluation of AODV, DSDV & DSR for Quasi Random Deployment of Sensor Nodes in Wireless Sensor Networks, International Conference on Devices and Communications (ICDeCom), pp. 1-5, [13] Wireless medium access control (MAC) and physical layer (PHY) specifications for low-rate wireless personal area networks (WPANs), IEEE Std (Revision of IEEE Std ), pp , Jul [14] C. Sergiou, V. Vassiliou, A. Paphitis, Hierarchical Tree Alternative Path (HTAP) algorithm for congestion control in wireless sensor networks, Ad Hoc Networks 11, pp , [15] C. Perkins, E. Royer, Ad-hoc on-demand distance vector routing, IEEE Workshop on Mobile Computing Systems and Applications, WMCSA, pp , [16] D. Maltz, J. Broch, J. Jetcheva, D. Johnson, The effects of on-demand behavior in routing protocols for multihop wireless ad hoc networks, IEEE Journal on Selected Areas in Communications 17, pp , [17] A. Gagarin, S. Hussain, L. Yang, Distributed hierarchical search for balanced energy consumption routing spanning trees in wireless sensor networks, Journal of Parallel and Distributed Computing 70, pp , Sep ACKNOWLEDGMENT This work has been supported by the Institute of Science and Technology of the Federal District (ICyTDF) under projects PICOO10-77 and 396/2010. REFERENCES [1] M. Yigitel, O. Incel, C. Ersoy, Qos-aware mac protocols for wireless sensor networks: A survey, Computer Networks 55, pp , Jun [2] W. Qiu, E. Skafidas, P. Hao, Enhanced tree routing for wireless sensor networks, Ad Hoc Networks 7, pp , May [3] M. Liu, S. Xu, S. Sun, An agent-assisted QoS-based routing algorithm for wireless sensor networks, Journal of Network and Computer Applications 35, pp , Jan [4] C. Mavromoustakis, H. Karatza, Optimized QoS priority routing for service tunability and overhead reduction using swarm based active network scheme, Computer Communications 29, pp , Mar [5] S. Santhi, G. Sadasivam, Performance Evaluation of Different Routing Protocols to Minimize Congestion in Heterogeneous Network, International Conference on Recent Trends in Information Technology (ICRTIT), pp , Jun [6] G. Ding, Z. Sahinoglu, P. Orlik, J. Zhang, B. Bhargava, Tree-Based Data Broadcast in IEEE and ZigBee Networks, IEEE Transactions on Mobile Computing 5, pp , [7] I. Demirkol, C. Ersoy, Energy and delay optimized contention for wireless sensor networks, Computer Networks 53, pp , Aug [8] F. Cuomo, E. Cipollone, A. Abbagnale, Performance analysis of IEEE wireless sensor networks: An insight into the topology formation process, Computer Networks 53, pp , Dec [9] W. Qiu, E. Skafidas, Q. Cheng, A hybrid routing protocol for wireless sensor networks, International Symposium on Communications and Information Technologies, ISCIT, pp , 2007.