SumMER Prosect REPORT ENERGY EFFICIENT ROUTING IN WIRELESS SENSOR NETWORK November 8, 2013 Under the guidence of Prof, Suchismitha C Mann Elappila Roll No: 212CS1089 National Institute Of Technology Rourkela Department Of Computer Science and Engineering,Abstract : Recently there has been a lot of interest in building and de- ploying sensor networks. These networks are composed of a high number of very simple nodes where most of them have to perform the function of a. router also. Energy consumption of these nodes is important as the power supplies of the node is provided by limited batteries, which circumseribe the lifetime of the links as well as whole networks. Therefore controlling the energy usage of network sensor node is critical for long lifespan of the network, Since the sensor nodes are acting like routers as well, the choice of routing algorithm would have a major role in the energy consumption control. As the well-known routing protocols use the optimized path, it will degrade the energy level of the nodes in that particular path. But instead energy efficient algorithms will overcome this problem. Keywords : Sensor nodes, Energy efficiency, Routing, Path Survivability factor.Contents Introduction Comparitive study of existing routing protocols 2.1 Routing Metrics 22 Routing Protocols 221 Flooding 22.2 Sensor Protocol for Information via Negatiation (SPIN) 22.3 Direct Diffusion 224 Heirarchical Routing 22.5 Proactive Routing Protocols 22.6 Reactive Routing Protocols Energy Efficient Routing Protocols 3.1, Energy Aware Routing 3.1.1 Setup Phase 3.1.2 Data Communication Phase 3.2. Sub-Game based Energy Aware Routing (SGEAR) 3.2.1 Setup Phase 32.2 Data Communication Phase 3.2.3. Route Maintenance Phase Conclusion References wae 10 10 13 13 uu 15 151 Introduction Wireless Sensor Networks (WSN) contain thousands of sensor nodes which has the abilities for sensing, computing, and communicati happening in its environment and the other interested elements, perform some simple computations and communicate amoung the peer nodes or directly to the base station, Deployement of the nodes of the sensor networks can be in the random fashion like dropping from aeroplane and sometimes it can be planed manually also, Because a sensor node needs to operate for a long time on a tiny battery, innovative techniques to eliminate nefficiencies that would short: ‘the main design goal of WSNs is to increase the lifetime of the network and prevent connectivity degradation by employing aggresive energy management techniques. the Position of the sensor nodes are usually not engineered or predetermined, and thus allows random deployment in inaccessible terrain or disaster relief operations. This implies ‘that the nodes are expected to perform sensing and communication with no continuous maintenance or human attendance and battery replacement. It limits the amount of energy available to the sensor nodes. Most WSN applications require large number of sensor nodes that cover large areas,necessitating an indirect (multi-hop) communication approach, ‘That is, sensor node must not only generate and disseminate their own information but also serves as router for forwarding packets. ‘The routing protocol designed for traditional networks cannot be used direetly in a sensor network because the sensor nodes should be self organizing, because of the ad hoc deployment and human unattendance and also because of drastically changing network topology. In sensor networks conservation of energy which is directly related to the network lifetime is considered relatively more important than the performance of the network in terms of quality of data sent. As the energy get depleted the network may be required to reduce the quality of the results in order to reduce the energy dissipation in ‘the nodes and thus lengthen the total network lifetime. ‘Therefore conservation of energy is considered to be more important than the performance of the network. In this paper I am comparing various existing routing protocols for wireless sensor networks and doing a detialed study about the energy efficient routing protocols and also putting forward a new modification for a dynamic game theoretical energy efficient routing protocol based on logarithmic path survivability factor. Each node will sense the events the lifetime of the network must be used. Hence of 2 Comparitive study of existing routing protocols ‘There are wide varieties of routing protocols for the communication between the sensor nodes of the WSN in the litrature. Some amongst them are Flooding, Data-Centric routing, Proactive routing, On-demand routing, Heirarchical routing etc. Routing protocol are responsible for identifying or discoverin to the intended reciever. On-demand or reactive protocols discover a route whenever a, source wants to send the data to a reciever and does not already have a route established, ‘This reactive route discovery makes some delay before the actual data transmission is routes form a source or senderoccuring, On the other hand proactive routing protocols discover or establish the route before they are actaully needed. These are also known as “table-driven’, because it uses a ng table that contains a list of destination combined with one or more next hope neighbors that lead towards these destinations and cost associated with each next hop option. Some protocols exhibit characteristics of both reactive and proactive protocols and belongs to hybrid routing protocol. ‘There are three different approaches through which the sensor data is collected. In tthe first approach ie, time-driven scheme nodes propagate their collected sensor data periodically to a sink. In the case of event-driven schemes nodes will only report their jerest occur. Finally, in query-driven scheme ‘the sink will request the data from the sensors when needed. It is the sink’s responsibility, to query the data. But the routing protocol is designed regardless of the scheme used collected information when the event of i in the network, based on the needs of the application and the resourees availabe in the network. 2.1 Routing Metrics Different routing metrics are used to express a variety of objectives of a routing protocol with respect to the consumption of these resources or the performance an application perceives. The commonly used routing metries in WSN are: © Minimum Hop: The most common metric used in routing protocols is minimum hop(or shortest hop), that is, the routing protocol attempts to find the path from. the sender to the destination that requires the smallest number of relay nodes (hops). However, since the minimum-hop approach does not consider the actual resource availability on each node, the resulting route is probably non-optimal in terms of delay, energy, and congestion avoidance. © Energy: Undoubtedly the most crucial aspect of routing in WSNs is energy efficiency. There are various different interpretations of energy efficiency, they are: ~ The mi amount of energy expended for the propagation of a single packet from the source to the destination. The total energy is then the sum of the energy consumed by each node along a route for receiving and transmitting the packet. mum energy consumed per packet : The goal is to minimize the total — Maximum time to network partition : The challenge is to reduce the energy consumption on nodes that are crucial to maintaining a network where every sensor node can be reached via at least one route. Premature expiration of those nodes should be prevented. — Minimum variance in node power levels : The challenge is to distribute the energy consumption across all nodes in the network as equally as possible. — Maximum (average) energy capacity : In this approach, the focus is less on the energy cost of packet propagation, but instead, ay capacity (ic., the current battery charge level) of the nodes. A routing protocol that the euses this metric would then favor r capacity from source to destination ates that have the largest total energy — Maximum minimum energy capacity : Here, instead of maximizing the energy capacities of the entire path, the primary routing goal could be to select the path with the largest minimum energy capacity. This technique also favors routes with larger energy reserves, but also protects low-capacity nodes from. premature expiration. © Quality-of-service: The term Quality-of-Service (QoS) refers to defined measures of performance in networks, including end-to-end latency (or delay) and throughput, but also jitter (variation in latency) and packet loss (or error rate). The different: routing protocols may be used to enjoy different quality of services y stable * Robustness: Many sensor applications may wish to use routes that st and reliable for long periods of time. However, this metric is rarely used alone. A routing protocol could identify several minimum-hop paths and then select the one with the highest total or average link quality along these paths. 2.2. Routing Protocols Different types of routing protocols which is there in the literature presently are, Flooding, SPIN family routing protocols, Direct diffusion, proactive routing protocols, on demand routing protocols, LEACH ete.. The comparison of each protocol is following one by one. 2.2.1 Flooding ‘The idea is to flood the data into the entire network. A sender node broadcasts packets to its immediate neighbors, which will repeat this process by rebroadeasting the packets to ‘their own neighbors until all nodes have received the packets or the packets have traveled for a maximum number of hops. With flooding, if there exists a path to the destination it, is guaranteed to receive the data. Flooding faces some challenges like Implosion, Overlap, Resource blindness. A node receiving a packet relays this packet to all its neighbors using, broadcasting, regardless of whether these neighbors have already received this packet from other neighbors. This leads to resource waste due to unnecessary transmit-and-receive operations which is known as Implosion. Sensors are often used to monitor overlapping ‘geographic areas, similar to the implosion problem, this also leads to resource waste since ‘the same information is sent twice to the same receiver. 2.2.2 Sensor Protocol for Information via Neg: n (SPIN) Sensor Protocols for Information via Negotiation (SPIN) is a family of negotiation-based, data-centric, and time-driven flooding protocols. To address the problems of implosion and overlap, SPIN nodes negotiate with their neighbors before they transmit data, al- lowing them to avoid unnecessary communications. SPIN is a family of protocols which work based on the negotiation and handshaking. SPIN uses meta-data to completelydiscribe the sensor node's collected data. SPIN-PP,SPIN-EC,[Link]-RL are the members of SPIN family protocols. SPIN-PP is for the networks which uses point-to-point transmission media, In SPIN-PP, data is flooded in three steps via a 3-way handshake protocol. First, when new data arrives, ‘a node advertises this event using an advertisement message (ADV) to its neighbors via ‘the data’s meta-data. Upon receiving an advertisement, a node checks whether it has al- ready received the described sensor data. If not, the node responds with a request for data (REQ) message, indicating its desire to receive the advertised data. Finally, the sender node responds to the REQ message with a DATA message, containing the advertised data. SPIN-EC is variation of this protocol, which adds a simple heuristic to add energy conservation to the SPIN-PP protocol. As long as all nodes have sufficient energy, they participate in the three-way handshake of the SPIN-PP protocol. A node replies to an advertisement only if it has sufficient energy to transmit the request and receive the re- quested data. Similarly, a node initiates the three-way handshake with its neighbors only if it believes that it can complete the protocol even if all neighbors request a copy of the data. SPIN-BC is designed for broadcast channels. All nodes within the hearing range of le will get the message. However, nodes must wait for transmission if the busy. Also, nodes do not immediately send out REQ message when they hear ‘the ADV message. Instead each node set a random timer and when this timer expires the node sends the REQ message. If, waiting for their timer to expire, other nodes are able to hear this message, they will stop their timers. This prevents sending the redundant copies of the same request, SPIN-RL is a reliable version of SPIN-BC, addressing packet loss and asymmetric ‘communications. First, each node keeps track of overheard REQ messages and if it does not receive a corresponding DATA message within a certain timeout interval, it assumes that either the REQ message or the DATA message did not arrive. In this case, the node rerequests the data by broadcasting an REQ message, specifying the identity of ‘a randomly selected node among the nodes that previously advertised this data in the message header. In addition, SPIN-RL limits the frequency with which DATA messages are sent out. That is, once a node sends a DATA message, it will wait a predetermined time before responding to any other requests for the same data, channel 2.2.3 Direct Diffusion ‘The main idea of directed diffusion is that nodes request data by sending interests for named data. This interest dissemination sets up gradients within the network that are used to direct sensor data toward the recipient. A sink node periodically broadcasts an interest message to its neighbors, which continue to broadcast the message throughout: ‘the network. Each node establishes a gradient toward the sink node, where a gradient is a reply link toward the neighbor from which the interest was received. Source will send the data back to the sink through this gradient. Direct diffusion is a query-based protocol,which may not be a good choice for certain sensor network applications, particularly where continuous data transfers are required. 2.2.4 Heirarchical Routing An example for heirarchical clustering algorithm is the Low-Energy Adaptive Clustering Hierarchy (LEACH) protocol, which combines a clustering approach with MAC-layer techniques. LEACH assumes that every cluster head can directly communieate with the base station. With LEACH, cluster heads are responsible for all communication between their cluster members and a base station and the aggregation of data coming from its cluster members in order to eliminate redundancies. LEACH can achieve significant energy savings and sensor nodes other than the cluster heads are not responsible for forwarding other node's data. 2.2.5 Proactive Routing Protocols Proactive routing protocols have the distinguishing characteristic of attempting to main- tain consistent up-to-date routing information from each node to every other node in the network. Every node maintains one or more routing tables that store the routing infor mation, and topology changes are propagated throughout the network as updates so that ‘the network view remains consistent. The protocols vary in the number of routing tables maintained and the method by which the routing updates are propagated. Two common. proactive protocols are discussed below briefly; there are many others in the literature also. ‘The Destination-Sequenced Distance-Vector Routing protocol (DSDV) is a example of proactive protocols. It is based on the Bellman-Ford algorithm for shortest paths and ensures that there is no loop in the routing tables. Every node in the network maintains the next hop and distance information to every other node in the network. Routing table updates are periodically transmitted throughout the network to maintain table consistency. ical Link-state Routing is a proactive protocol in which each node floods the cost of all ‘the links to which it is connected throughout the network. Every node then works out the cost of reaching every other node using shortest path algorithms. In addition, the protocol works correctly even if unidirectional links are present, whereas DSDV assumes bidirectional links. 2.2.6 Reactive Routing Protocols In contrast to proactive routing protocols, reactive protocols create routes only when desired. This means that an explicit route discovery process creates routes and this is ini- tiated only on an as-needed basis. It can be either sourceinitiated or destination-initiated, Source-initiated routing means that it is the source node that begins the discovery process, while destination-initiated is the opposite. Once a route has been established, the routediscovery process ends, and a maintenance procedure preserves it until the route breaks down or is no longer desired. Ad-Hoc On-Demand Distance Vector Routing (AODV) is a routing protocol also based on the distance vector algorithm like DSDV, but the difference is that AODV is reactive. It is a source-initiated protocol, with the source node broadcasting a Route Request (RREQ) when it determines that it needs a route to a destination and does not have one available. This request is broadcast till the destination or an intermediate node with a, fresh enough route to the destination is located. Intermediate nodes record the address of the neighbor from which the first copy of the broadcast packet is received in their route tables, thus establishing a reverse path. Dynamic Source Routing (DSR) is a reactive protocol that is also source-initiated and is based on the concept of source routing, i.e. the source specifies the entire route to be taken by a packet, rather than just the next hop. If the source node does not have a route, it floods the network with a Route Request (RREQ). Any node that has a path to ‘the destination can reply with a Route Reply (RREP) to the source. This reply contains ‘the entire path recorded in the RREQ packet. The entire path is added to the header of every packet to the destination, thus it is called source routing, 3 Energy Efficient Routing Protocols ‘The potential problem in current protocols is that they find the lowest energy route and use that for every communication. However, that is not the best thing to do for network lifetime, Using a low energy path frequently leads to energy depletion of the nodes along, ‘that path and in the worst case may lead to network partition. 3.1 Energy Aware Routing ‘The basic idea of this Energy Aware Routing Protocol ([Link] et al.) is to inerease the survivability of networks, it may be necessary to use sub-optimal paths occasionally. This ensures that the optimal path does not get depleted and the network degrades gracefully ‘as a whole rather than getting partitioned. Multiple paths are found between source and destinations, and each path is assigned a probability of being chosen, depending on the energy metric. Every time data is to be sent from the source to destination, one of the paths is randomly chosen depending on the probabilities. This means that none of the paths is used all the time, preventing energy depletion. Also different paths are tried continously, improving tolerance to nodes moving around the network. this is a reactive routing protocol and also a destination-initiated protocol where the consumer of data initiates the route request and maintains the route subsequently. The protocol has three phases: © Setup phase or interest propagation - Localized flooding occurs to find all the routesfrom source to destination and their energy costs. This is when routing (interest) tables are built up. © Data Comn destination mication phase or data propagation - Data is sent from source to using the information from the earlier phase. This is when paths are chosen probabilistically according to the energy costs that were calculated earlier. # Route maintenance - Route maintenance is minimal. Localized flooding is performed infrequently from destination to source to keep all the paths alive. 3.1.1 Setup Phase 1. The destination node initiates the connecti n by flooding the network in the direction of the source node. It also sets the Cost field to zero before sending the request. Cost(Np) =0 2. Every intermediate node forwards the request only to the neighbors that are closer to the source node than oneself and farther away from the destination node. Thus at anode 1Vj, the request is sent only to a neighbor Vj which satisfies: (Ni, Ns) > dCN5,No) d(N;, Np) < a(Nj,Np) where d(N;, Nj) is the distance between the N and Nj 3. On receiving the request, the energy metric for the neighbor that sent the request: is computed and is added to the total cost of the path. Thus, if the request is sent from node Nj to node Nj, Nj calculates the cost of the path as: On, x, = Cost(Ni) + Metric(Nj, Ni) 4. Paths that have a very high cost are discarded and not added to the forwarding table. Only the neighbors Nj with paths of low cost are added to the forwarding, table FT; of Nj FT; = {iJCn,.n, a1;as > a3), the energy consumption of one hop of node $ be cs. According to the above protocol, when it comes to the choice of M, M chooses the path from L1 and Ly that satisfy Ma:(a,/c1,a2/e2), when it comes to the path choice of 8, S chooses the path between 4 and L) that satisfy maz(ay/(e1 + ¢s),@2/(c2 + ¢s)). Unfortunately there exist a scenario that satisfies (ai/e1) < (az/e2) and (ai/(c1 +s) > a2/(c2 + es), for example, a) = 3,¢) = 5,@) = 2,c) = 3,¢, =2. Here, 0.66 3 5 2 3 Figure 1: Irregular change in path choosing. u‘That means the node M will choose path L2 while node $ will choose path Li. In order to overcome this irregular changing scenario here it is putting forward a modification for the simple path survivability factor. Let ¢ be the totle energy consumption of the path L, let a be the minimum power available value amoung the nodes in path L, define f(a,e) as the generalized path survivability factor of L which satisfies © fis a monotone increasing function of a; © fis a monotone decreasing function of e. ‘The generalized path survivability factor is an extension of path survivability factor. ‘To optimize the lifetime of the network, the total energy consumption along the path and the bottleneck node because of less energy should both be considered by relaying nodes, so the generalized survivability factor is defined to give a trade off between these ‘two factors. Logarithm is such a monotone increasing function and so at place of generalized path survivability factor, it can get replaced with the logarithmic path survivability factor. That is, ‘This logarithmic path survivability factor will not gain the irregular change when some new node is added to the path. That is, for the preveous example, tog3 1 = FE = 0008 tog? ta = = 0003 , _ tog3 1, = GE 0.50 1, = 82 Lo, jogs ‘That means both the nodes will choose the same path. So this proposed scheme will work in three phases: Setup phase, Data Communication Phase, and Route Maintenance Phase. 123.2.1 Setup Phase 1. Sink node initiates a broadcasting of interest along with the path choosing factor (a,¢) 2. Once receiving the interest message, each intermediate node i adds the path choos- ing factors of the down direction nodes into the routing table RT; RT, = {jl f(ajs¢5) > a= min([Link])),a> 1, teRT} where a is a constant, which can adjust the size of the routing table. 3. intermediate node i calculates its logarithmic survivability factors, choose the bigger one as routing choice of node # while communicating: D(i) = {il f(aj,¢5) = marl flat, ex)), teRT} 4. intermediate node i calculate its path coosing factor according to the path choosing factor of D(i) nin([Link]() cj = epi) + Metrie(i, D(i)) ‘hich Metric(i, D(i)) is the energy consumption between node i and node D(i); 5. intermediate node i replaces the path choosing factor of the interest with the calculated path choosing factor of Node i and forward the interest to up-direction nodes. 3.2.2 Data Communication Phase L. once the interest message is forwarded to the source node, the source node makes routing choice according to the logarithmic survivabili path choosing factors; factor calculated with the 2. each intermediate node makes the routing choice according to the logarithmic path survivability factor calculated with the path choosing factors when relaying. 13LL sink initiates broadcasting of interest tee the effectiveness of each path; yeles, restart Setup Phase to guar 2. a threshold k is given to each intermediate node. Once the survivability factor of the node is below the threshold, intermediate nodes can trigger Setup Phase by sending a restart message to Sink node. u4 Conclusion ‘The proposed scche is almost same as the energy aware routing,except that it forms the routing table by the logarithmic path survivability factors and chooses routing paths according to those survivability factors. Because it makes the choices of routing paths in ‘determined manner rather than generated randomly, routing will be more stable wuthin ‘cycle and this feature will be more predictable. Therefore this can be easly combined with technologies of sleep scheduling. however the legth of a cycle is a key point to this scheme. If the length is too short it will degenerate to Energy Aware routing, on the other hand if it is too high to reach the same path will be choosen for longer time and may lead to the network disconnection. So the value of k should be well set so that it can be used in combination with the sleep scheduling techniques and get energy efficient results, 5 References (1) Dayang Sun, Xuan Huang, Yanheng Liu, and Hui Zhong. Predictable Energy Aware Routing based on Dynamic Game Theory in WSN. In: SciVerse ScienceDirect Computer and Electrical Engineering 39 (2013) 1601-1608. (2) Shah R, Rabaey J. Energy aware routing for low energy ad hoc networks. In: Proc of the IEEE wireless communication and networking conf Orlando: IEEE, ‘Communication Society; 2002. p. 350-55, (3) Julius Hossain M, Chae Oksam, Mamun-or-Rashid Md, Seon Hong Choong. Costeffec- tive maximum lifetime routing protocol for wireless sensor networks. In: Proceedings of the advanced industrial conference on telecommunication/service assurance with partial and intermittent resources conference /E-learning on telecommunication workshop. IEEE Computer Society; 2005. p. 314-19, (4) Imad Mahgonb, Mohammad Ilyas Sensor Network Protocols, CRC taylor and francis, Group;2006. (5) Waltenegus Dargie, Christian Poellabaner Fundamentals of Wireless Sonsor Networks John Wiely and sons Ltd. 2010 15