Network structure routing_protocols_in_wsn


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Network structure routing_protocols_in_wsn

  1. 1. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012 Network Structure Routing Protocols in WSN Jain Pratik P1, Manjunath CR2 , DrNagaraj GS3 1 nd MTech 2 Sem, SET, JAIN University, 2 Asst Prof, Dept of CSE, SET, JAIN University, 3 professor, Dept of CSE, RVCEABSTRACT- Routing protocols are in charge of play different roles in the network. In location-discovering and maintaining the routes in the based routing, sensor nodes positions are exploitednetwork. This paper presents a review of the main to route data in the network [2]. A routing protocolrouting protocols proposed for Network Structure in is considered adaptive if certain system parametersWSN. Based on the study, they are classified into can be controlled in order to adapt to the currentthree categories: SPIN & Directed Diffusion in Flat-based, LEACH, PEGASIS & SOP in Hierarchical- network conditions and available energy levels.based, SPAN & GOFAR in Location-based, which Another class of routing protocols is called theunderlying Network Structure. cooperative routing protocols. In cooperative routing, nodes send data to a central node whereKeywords- Wireless Sensor Network, Routing data can be aggregated and may be subject toProtocols: Flat-based, Hierarchical-based, Location- further processing, hence reducing route cost inbased. terms of energy use [2]. I. INTRODUCTION II. Flat-based Routing Protocols Wireless Sensor Networks (WSN) are The first category of routing protocol isintended for monitoring an environment. The main the multi-hop flat routing protocols. In flattask of a wireless sensor node is to sense and networks, each node typically plays the same rolecollect data from a certain domain, process them and sensor nodes collaborate together to performand transmit it to the sink where the application the sensing task. Due to the large number of suchlies. Areas of applications are like, physical nodes, it is not feasible to assign a global identifierparameters (e.g., humidity, temperature, light, etc.), to each node. This consideration has led to datato safety and security-oriented (intrusion detection, centric routing, where the BS sends queries tonatural disaster detection, etc.), to monitor soldiers certain regions and waits for data from the sensorsin the battle field (military applications), patients in located in the selected regions. Since data is beingnursing institutes (e-health applications), fire requested through queries, attribute-based namingbrigades and policemen (security/safety is necessary to specify the properties of data. Earlyapplications) [1, 6]. works on data centric routing, e.g., SPIN and The use of potentially unique identifier directed diffusion were shown to save energysuch as the MAC (Medium Access Control) through data negotiation and elimination ofaddress or the GPS coordinates is not redundant data.recommended as it forces a significant payload inthe messages. However, this drawback is easilyovercome in wireless sensor networks since an IPaddress is not required to identify the destinationnode of a specific packet. In fact, attribute-basedaddressing fits better with the specificities ofwireless sensor networks. In this case, an attributesuch as node location and sensor type is used toidentify the final destination. Once nodes areidentified, routing protocols are in charge ofconstructing and maintaining routes betweendistant nodes. The different ways in which routingprotocols operate make them appropriate forcertain applications [6]. Figure1: Routing Protocols in WSN In general, network structure routing inWSNs can be divided into flat-based routing, SPIN:hierarchical-based routing, and location-based Sensor Protocols for Information viarouting depending on the envirnoment. Negotiation (SPIN) that enables a user to query any In flat-based routing, all nodes are node and get the required information immediately.typically assigned equal roles or functionality. In These protocols make use of the property thathierarchical-based routing, however, nodes will nodes in close proximity have similar data, and 23
  2. 2. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012hence there is a need to only distribute the data that SPIN-RL: When a channel is lossy, a protocolother nodes do not posses. The SPIN family of called SPIN-RL is used where adjustments areprotocols uses data negotiation and resource- added to the SPIN-PP protocol to account for theadaptive algorithms. Nodes running SPIN assign a lossy channel.high-level name to completely describe their Advantages: Topological changes are localizedcollected data (called meta-data) and perform meta- since each node needs to know only its single-hopdata negotiations before any data is transmitted. neighbors. SPIN provides much energy savingsThis assures that there is no redundant data sent than flooding and meta-data negotiation almostthroughout the network. The semantics of the meta halves the redundant data. However, SPINs datadata format is application-specific and is not advertisement mechanism cannot guarantee thespecified in SPIN. For example, sensors might use delivery of data. To see this, consider thetheir unique IDs to report meta-data if they cover a application of intrusion detection where datacertain known region. In addition, SPIN has access should be reliably reported over periodic intervalsto the current energy level of the node and adapts and assume that nodes interested in the data arethe protocol it is running based on how much located far away from the source node and theenergy is remaining. These protocols work in a nodes between source and destination nodes are nottime-driven fashion and distribute the information interested in that data, such data will not beall over the network, even when a user does not delivered to the destination at all [2].request any data. The SPIN family is designed toaddress the deficiencies of classic flooding by Directed Diffusion:negotiation and resource adaptation. The SPIN Directed diffusion is a data-centric (DC)family of protocols is designed based on two basic and application-aware paradigm in the sense thatideas: all data generated by sensor nodes is named by 1) Sensor nodes operate more efficiently and attribute-value pairs. The main idea of the DCconserve energy by sending data that describe the paradigm is to combine the data coming fromsensor data instead of sending all the data; for different sources enroute (in-network aggregation)example, image and sensor nodes must monitor the by eliminating redundancy, minimizing the numberchanges in their energy resources. of transmissions; thus saving network energy and 2) Conventional protocols like flooding or prolonging its lifetime. Unlike traditional end-to-gossiping based routing protocols waste energy and end routing, DC routing finds routes from multiplebandwidth when sending extra and un-necessary sources to a single destination that allows in-copies of data by sensors covering overlapping network consolidation of redundant data.areas. The drawbacks of flooding include In directed diffusion, sensors measureimplosion, which is caused by duplicate messages events and create gradients of information in theirsent to the same node, overlap when two nodes respective neighborhoods. The base stationsensing the same region will send similar packets requests data by broadcasting interests. Interestto the same neighbor and resource blindness by describes a task required to be done by theconsuming large amounts of energy without network. Interest diffuses through the network hop-consideration for the energy constraints. by-hop, and is broad-cast by each node to its SPIN is a 3-stage protocol as sensor nodes neighbors. As the interest is propagated throughoutuse three types of messages ADV, REQ and DATA the network, gradients are setup to draw datato communicate. ADV is used to advertise new satisfying the query towards the requesting node,data, REQ to request data, and DATA is the actual i.e., a BS may query for data by disseminatingmessage itself. The protocol starts when a SPIN interests and intermediate nodes propagate thesenode obtains new data that it is willing to share. It interests. Each sensor that receives the interestdoes so by broadcasting an ADV message setup a gradient toward the sensor nodes fromcontaining meta-data. If a neighbor is interested in which it receives the interest. This processthe data, it sends a REQ message for the DATA continues until gradients are setup from the sourcesand the DATA is sent to this neighbor node. The back to the BS. More generally, a gradient specifiesneighbor sensor node then repeats this process with an attribute value and a direction. The strength ofits neighbors. As a result, the entire sensor area will the gradient may be different towards differentreceive a copy of the data. neighbors resulting in different amounts of The SPIN family of protocols includes information flow. At this stage, loops are notmany protocols. checked, but are removed at a later stage. Figure2SPIN-BC: This protocol is designed for broadcast shows an example of the working of directedchannels. diffusion ((a) sending interests, (b) buildingSPIN-PP: This protocol is designed for a point to gradients, and (c) data dissemination). Whenpoint communication, i.e., hop-by-hop routing. interests fit gradients, paths of information flow areSPIN-EC: This protocol works similar to SPIN- formed from multiple paths and then the best pathsPP, but with an energy heuristic added to it. are reinforced so as to prevent further flooding 24
  3. 3. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012according to a local rule. In order to reduce known techniques with special advantages relatedcommunication costs, data is aggregated on the to scalability and efficient communication. Asway. The goal is to find a good aggregation tree such, the concept of hierarchical routing is alsowhich gets the data from source nodes to the BS. utilized to perform energy-efficient routing inThe BS periodically refreshes and re-sends the WSNs. In a hierarchical architecture, higher energyinterest when it starts to receive data from the nodes can be used to process and send thesource(s). This is necessary because interests are information while low energy nodes can be used tonot reliably transmitted throughout the network. perform the sensing in the proximity of the target. This means that creation of clusters and assigning special tasks to cluster heads can greatly contribute to overall system scalability, lifetime, and energy efficiency. Hierarchical routing is an efficient way to lower energy consumption within a cluster and by performing data aggregation and fusion in order to decrease the number of transmitted messages to the BS. Hierarchical routing is mainly two-layer routing where one layer is used to select cluster heads and the other layer is used for routing. LEACH protocol: Low Energy Adaptive Clustering Hierarchy (LEACH) is a cluster-based protocol, which includes distributed cluster formation. LEACH, randomly selects a few sensor nodes as cluster-heads (CHs) and rotate this role to evenly distribute the energy load among the sensors in the network. In LEACH, the cluster-head (CH) nodes compress data arriving from nodes that belong to the respective cluster, and send an aggregated packet to the base station in order to reduce the amount of information that must be transmitted to the base station. LEACH uses a TDMA/CDMA MAC to reduce inter-cluster and intra-cluster collisions. However, data collection is centralized and is performed periodically. Therefore, this protocol is most appropriate when there is a needFigure2: Examples of Interest Diffusion in Sensor for constant monitoring by the sensor network. A Network user may not need all the data immediately. Hence, periodic data transmissions are unnecessary which Directed diffusion differs from SPIN in may drain the limited energy of the sensor nodes.two aspects: After a given interval of time, a randomized 1) Directed diffusion issues on demand rotation of the role of the CH is conducted so thatdata queries as the BS send queries to the sensor uniform energy dissipation in the sensor network isnodes by flooding some tasks. In SPIN, however, obtained. The authors found, based on theirsensors advertise the availability of data allowing simulation model, that only 5% of the nodes needinterested nodes to query that data. to act as cluster heads. 2) All communication in directed The operation of LEACH is separated intodiffusion is neighbor-to-neighbor with each node two phases, the setup phase and the steady statehaving the capability of performing data phase.aggregation and caching. Unlike SPIN, there is no In the setup phase, the clusters areneed to maintain global network topology in organized and CHs are selected. In the steady statedirected diffusion. However, directed diffusion phase, the actual data transfer to the base stationmay not be applied to applications (e.g., takes place. The duration of the steady state phaseenvironmental monitoring) that require continuous is longer than the duration of the setup phase indata delivery to the BS. order to minimize overhead. During the setup phase, a predetermined fraction of nodes, p, electIII. Hierarchical-based Routing Protocols themselves as CHs as follows. A sensor node chooses a random number, r, between 0 and 1. If Hierarchical or cluster-based routing, this random number is less than a threshold value,originally proposed in wireline networks, are well- T(n), the node becomes a cluster-head for the 25
  4. 4. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012current round. The threshold value is calculated fashion. Simulation results showed that PEGASISbased on an equation that incorporates the desired is able to increase the lifetime of the network twicepercentage to become a cluster-head, the current as much the lifetime of the network under theround, and the set of nodes that have not been LEACH protocol. Such performance gain isselected as a cluster-head in the last (1/P) rounds, achieved through the elimination of the overheaddenoted by caused by dynamic cluster formation in LEACHG. It is given by: and through decreasing the number of transmissions and reception by using data aggregation. Where, G is the set of nodes that are SOP:involved in the CH election. Each elected CH Self Organizing Protocol is used to buildbroadcast an advertisement message to the rest of architecture to support heterogeneous sensors.the nodes in the network that they are the new These sensors can be mobile or stationary. Somecluster-heads. All the non-cluster head nodes, after sensors probe the environment and forward the datareceiving this advertisement, decide on the cluster to a designated set of nodes that act as which they want to belong to. This decision is Router nodes are stationary and form the backbonebased on the signal strength of the advertisement. for communication. Collected data are forwardedThe non cluster-head nodes inform the appropriate through the routers to the more powerful BS nodes.cluster-heads that they will be a member of the Each sensing node should be able to reach a routercluster. After receiving all the messages from the in order to be part of the network. A routingnodes that would like to be included in the cluster architecture that requires addressing of each sensorand based on the number of nodes in the cluster, node has been proposed. Sensing nodes arethe cluster-head node creates a TDMA schedule identifiable through the address of the router nodeand assigns each node a time slot when it can they are connected to. The routing architecture istransmit. This schedule is broadcast to all the nodes hierarchical where groups of nodes are formed andin the cluster. [2]. merge when needed. Local Markov Loops (LML) algorithm, which performs a random walk onPEGASIS: spanning trees of a graph, was used to support fault Power-Efficient Gathering in Sensor tolerance and as a means of broadcasting. SuchInformation Systems (PEGASIS) is a near optimal approach is similar to the idea of virtual grid usedchain-based protocol. The basic idea of the in some other protocols that will be discussed laterprotocol is that in order to extend network lifetime, under location-based routing protocols. In thisnodes need only communicate with their closest approach, sensor nodes can be addressedneighbors and they take turns in communicating individually in the routing architecture, and hencewith the base-station. When, the round of all nodes it is suitable for applications where communicationcommunicating with the base-station ends, a new to a particular node is required. Furthermore, thisround will start and so on. This reduces the power algorithm incurs a small cost for maintainingrequired to transmit data per round as the power routing tables and keeping a balanced routingdraining is spread uniformly over all nodes. Hence, hierarchy. It was also found that the energyPEGASIS has two main objectives: consumed for broadcasting a message is less than 1) Increase the lifetime of each node by that consumed in the SPIN protocol. This protocol,using collaborative techniques and as a result the however, is not an on-demand protocol especiallynetwork lifetime will be increased. in the organization phase of algorithm. Another 2) Allow only local coordination between issue is related to the formation of hierarchy. Itnodes that are close together so that the bandwidth could happen that there are many cuts in theconsumed in communication is reduced. Unlike network, and hence the probability of applyingLEACH, PEGASIS avoids cluster formation and reorganization phase increases, which will be anuses only one node in a chain to transmit to the BS expensive operation [2].instead of using multiple nodes. To locate the closest neighbor node in VGA:PEGASIS, each node uses the signal strength to Virtual Grid Architecture routing is anmeasure the distance to all neighboring nodes and energy-efficient routing paradigm, proposed inthen adjusts the signal strength so that only one utilizes data aggregation and in-network processingnode can be heard. The chain in PEGASIS will to maximize the network lifetime. A GPS-freeconsist of those nodes that are closest to each other approach is used to build clusters that are fixed,and form a path to the base-station. The aggregated equal, adjacent, and non-overlapping withform of the data will be sent to the base-station by symmetric shapes. In square clusters were used toany node in the chain and the nodes in the chain obtain a fixed rectilinear virtual topology. Insidewill take turns in sending to the base-station. The each zone, a node is optimally selected to act aschain construction is performed in a greedy 26
  5. 5. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012cluster-head. Data aggregation is performed at two certain region rather than sending the interests tolevels: local and then global. The set of cluster- the whole network. By doing this, GEAR canheads, also called Local Aggregators (LAs), conserve more energy than directed diffusion. Eachperform the local aggregation, while a subset of node in GEAR keeps an estimated cost and athese Las, are used to perform global aggregation. learning cost of reaching the destination through itsHowever, the determination of an optimal selection neighbors. The estimated cost is a combination ofof global aggregation points, called Master residual energy and distance to destination. TheAggregators (MAs), is NP-hard problem. Figure 6 learned cost is a refinement of the estimated costillustrates an example of fixed zoning and the that accounts for routing around holes in theresulting VGA used to perform two level data network. A hole occurs when a node does not haveaggregation. Note that the location of the base any closer neighbor to the target region than itself.station is not necessarily at the extreme corner of If there are no holes, the estimated cost is equal tothe grid rather it can be located at any arbitrary the learned cost. The learned cost is propagated oneplace. hop back every time a packet reaches the destination so that route setup for next packet will be adjusted. There are two phases in the algorithm: 1) Forwarding packets towards the target region: Upon receiving a packet, a node checks its neighbors to see if there is one neighbor, which is closer to the target region than itself. If there is more than one, the nearest neighbor to the target region is selected as the next hop. If they are all further than the node itself, this means there is a hole. In this case, one of the neighbors is picked to forward the packet based on the learning cost function. This choice can then be updatedFigure3: Regular shape tessellation applied to the according to the convergence of the learned costnetwork area. In each zone, a cluster-head is during the delivery of packets.selected for local aggregation. Subsets of those 2) Forwarding the packets within thecluster-heads, called Master nodes, are optimally region: If the packet has reached the region, it canselected to do global aggregation. be diffused in that region by either recursive geographic forwarding or restricted flooding. Restricted flooding is good when the sensors are IV. Location based routing protocols not densely deployed. In high-density networks, In this kind of routing, sensor nodes are recursive geographic flooding is more energyaddressed by means of their locations. The distance efficient than restricted flooding. In that case, thebetween neighboring nodes can be estimated on the region is divided into four sub regions and fourbasis of incoming signal strengths. Relative copies of the packet are created. This splitting andcoordinates of neighboring nodes can be obtained forwarding process continues until the regions withby exchanging such information between only one node are left [2].neighbors. Alternatively, the location of nodes maybe available directly by communicating with a GOAFR:satellite, using GPS (Global Positioning System), if The Greedy Other Adaptive Face Routing.nodes are equipped with a small low power GPS The greedy algorithm of GOAFR always picks thereceiver. To save energy, some location based neighbor closest to a node to be next node forschemes demand that nodes should go to sleep if routing. However, it can be easily stuck at somethere is no activity. More energy savings can be local minimum, i.e. no neighbor is closer to a nodeobtained by having as many sleeping nodes in the than the current node. Other Face Routing (OFR) isnetwork as possible. The problem of designing a variant of Face Routing (FR). The Face Routingsleep period schedules for each node in a localized (FR) algorithm is the first one that guaranteesmanner was addressed in. In the rest of this section, success if the source and the destination arewe review most of the location or geographic based connected. However, the worst-case cost of FR isrouting protocols. proportional to the size of the network in terms of number of nodes. The first algorithm that canGEAR: The protocol, called Geographic and compete with the best route in the worst-case is theEnergy Aware Routing (GEAR), uses energy aware Adaptive Face Routing (AFR) algorithm.and geographically-informed neighbor selection Moreover, by a lower bound argument, AFR isheuristics to route a packet towards the destination shown to be asymptotically worst-case optimal. Butregion. The key idea is to restrict the number of AFR is not average-case efficient. OFR utilizes theinterests in directed diffusion by only considering a face structure of planar graphs such that the 27
  6. 6. National Conference on Current Trends in Computer Science and Engineering - CSECONF2012 nmessage is routed from node s to node t by protocols in wireless sensor networks which havetraversing a series of face boundaries. The aim is to been presented in the literature.find the best node on the boundary, i.e., the closest Overall, the routing techniques are enode to the destination t by using geometric planes. classified based on the network structure into threeWhen finished, the algorithm returns to s the best categories: SPIN & Directed Diffusion in Flat Flat-node on the boundary. The simple greedy based, LEACH, PEGASIS & SOP in Hierarchical Hierarchical-algorithm behaves well in dense networks, but it based, SPAN & GOFAR in Location Location-based.fails for very simple configurations. It was shownthat GOAFR algorithm can achieve both worst worst- REFERENCES [1] Luca Borsani, Sergio Guglielmi, Alessandro Redondi &case optimality and average-case efficiency. Based case Matteo Cesana, “Tree based Routing protocols in WSN,”on the simulation results of GOAFR, there are Dipartimento di Elettronica e Informazio Informazione,Politecnico diseveral ways to further improve the average case average- Milano, Italyperformance. It was also shown that GOAFR [2] Jamal N. Al-Karaki, Ahmed E. Kamal, “ Karaki, “Routing Techniques in Wireless Sensor Networks: A Survey*,” Dept. of Electricaloutperforms other prominent algorithms, such as and Computer EngineeringGPSR or AFR [2]. Iowa State University, Ames, Iowa [3] M.Simek, D.Komosny, R.Burget & J.S.Silva, Multicast J.S.Silva,”SPAN: Routing in Wireless Sensor Networks Networks,” Member, IEEE, Department of Telecommunications & Department of Another position based algorithm called Informatics EngineeringSPAN selects some nodes as coordinators based on [4] Rong Fan, Jian Chen, Jian-Qing Fu, Ling Qing Ling-Di Ping,” Atheir positions. The coordinators form a network Steiner-Based Secure Multicast Routing Protocol for Wireless Basedbackbone that is used to forward messages. A node Sensor Network,” College of Computer Science and Technology, Zhejiang University, presenting on 2010 Second ,should become a coordinator if two neighbor of a neighbors International Conference on Future Networks Hang Zhou, Networks,non-coordinator node cannot reach each other coordinator Chinadirectly or via one or two coordinators (3 hop [5] Shibo Wu, K. Selc¸uk Candan,” GMP: Distributedreachability). New and existing coordinators are Geographic Multicast Routing in Wireless Sensor Networks Networks,” Department of Computer Science, Arizona State University,not necessarily neighbors, which, makes the design Tempe, AZ, USA.less energy efficient because of the need to [6] Luis Javier García Villalba *, Ana Lucila Sandoval Orozco,maintain the positions of two or three hop Alicia Triviño Cabrera and Cláudia Jacy Barenco Abbas,” cianeighbors in the complicated SPAN algorithm [2]. Routing Protocols in Wireless Sensor Networks”, Department of Software Engineering and Artificial Intelligence, Ciudad Universitaria, Madrid, Spain. Figure 4: Classification and Comparision of omparision Routing Protocols in WSN. V. CONCLUSION Routing in sensor networks is a new a areaof research and rapidly growing set of researchresults. In this paper, we presented acomprehensive survey of network structure routing 28