Routing in a Wireless Sensor Network A project report submitted toThe Department of Computer Science and Engineering, Indian School Mines Dhanbad In partial fulfillment of the requirement for the award of degree of Bachelor of Technology in Computer Science and Engineering By Mohammad Kafee Uddin (2009JE0651) Aloukik Mishra(2009JE0640) Under the guidance of Prof P. K. Jana Dept. of Computer Science and Engg ISM, Dhanbad
Department of Computer Science and Engineering Indian School of Mines, Dhanbad-826004 Department of Computer Science and Engineering Indian School of Mines, Dhanbad Dated: 3-12-2012 CERTIFICATEThis is to certify that the dissertation entitled “ Routing in a Wireless SensorNetwork” is being submitted to Indian School of Mines, Dhanbad by MohammadKafee Uddin(2009Je0651) and Aloukik Mishra(2009JE0640) and in partialfulfillment of their Bachelor of Technology degree in Computer Science &Engineering of the same institution incorporates the result of his own work,carried out under my supervision and guidance. This dissertation has not beensubmitted for any other degree, elsewhere to the best of my knowledge. Prof P. K. Jana HOD, CSE ISM Dhanbad
AbstractRecent advances in wireless sensor networks have led to many newprotocols specifically designed for sensor networks where energyawareness is an essential consideration. Most of the attention, however,has been given to the routing protocols since they might differ dependingon the application and network architecture. This paper surveys recentrouting protocols for sensor networks and presents a classification for thevarious approaches pursued. We have classified routing protocolsaccording to three different parameters, namely Mode of Functioning andType of Target Applications, Participation style of the Nodes, and theNetwork Structure. Each routing protocol is described and discussed underthe appropriate category.
Acknowledgement First and foremost, we would like to express our sincere gratitudeto our Project guide, Prof P. K. Jana, Head of the Department ofComputer Science and Engineering for providing us with such anopportunity and permitting this project in the department. We were privileged to experience a sustained and involvedinterest on the part of our mentors. This encouraged us to boldly stepinto what was an unexplored expanse before us. Exploring new realmsin the field of Computer Science and Engineering helped unveil newinsights into the field expanding our knowledge of the same. We would also like to thank our friends who were ready with apositive comment all the time, whether it was an off-hand comment toencourage us or a constructive piece of criticism. Last but not least, we would like to thank the faculty membersand the Department, in general, for extending a helping hand at everyjuncture of need.Mohammad Kafee Uddin Aloukik MishraLocation: Location:Date: Date:
ContentsCertificateAbstractAcknowledgements1 Introduction to WSN2 Classification Of Routing Protocols2.1 Based on Mode of Functioning and Type of Target Applications 2.1.1 Proactive :- 2.1.2 Reactive :- 2.1.3 Hybrid :-2.2 According to the Participation style of the Nodes 2.2.1 Direct Communication :- 2.2.2 Flat :- 2.2.3 Clustering Protocols :- 2.3 Depending on the Network Structure 2.3.1 Data Centric :- 2.3.2 Hierarchical :- 2.3.3 Location Based :-3 Data Dissemination Protocols 3.1 Flooding 3.2 Gossiping 3.3 Rumor Routing 3.4 Sequential Assignment Routing 3.5 Direct Diffusion 3.6 Sensor Protocol for Information via Negotiation 3.7 Geographic Hash Table4 Data Gathering Protocols 4.1 Direct Transmission 4.2 Power Efficient Gathering for Sensor Information Systems 4.3 LEACH
1. Introduction to WSNRecent advances in micro-electro-mechanical systems and low power andhighly integrated digital electronics have led to the development of micro-sensors. Such sensors are generally equipped with data processing andcommunication capabilities. The sensing circuitry measures the ambientconditions related to the environment surrounding the sensor andtransforms them into an electric signal. Processing such a signal revealssome properties about objects located and/or events happening in thevicinity of the sensor. The sensor sends such collected data, usually viaradio transmitter, to a command center (sink) either directly or through adata concentration center (a gateway). The decrease in the size and cost ofsensors, resulting from such technological advances, has fueled interest inthe possible use of large set of disposable unattended sensors. Suchinterest has motivated intensive research in the past few years addressingthe potential of collaboration among sensors in data gathering andprocessing and the coordination and management of the sensing activityand data flow to the sink. A natural architecture for such collaborativedistributed sensors is a network with wireless links that can be formedamong the sensors in an ad hoc manner. A general sensor node is madeup of four basic components as shown in Fig. 1: a sensing unit, aprocessing unit, a transceiver unit and a power unit. They may also haveapplication dependent additional components such as a location findingsystem, a power generator and a mobilizer. Sensing units are usuallycomposed of two subunits: sensors and analog to digital converters(ADCs). The analog signals produced by the sensors based on theobserved phenomenon are converted to digital signals by the ADC, andthen fed into the processing unit. The processing unit, which is generallyassociated with a small storage unit, manages the procedures that makethe sensor node collaborate with the other nodes to carry out the assignedsensing tasks. A transceiver unit connects the node to the network. One ofthe most important components of a sensor node is the power unit. Powerunits may be supported by a power scavenging unit such as solar cells.There are also other subunits, which are application dependent. Most of thesensor network routing techniques and sensing tasks require theknowledge of location with high accuracy. Thus, it is common that a sensornode has a location finding system. A mobilizer may sometimes be neededto move sensor nodes when it is required to carry out the assigned tasks.
Networking unattended sensor nodes are expected to have significantimpact on the efficiency of many military and civil applications such ascombat field surveillance, security and disaster management. Thesesystems process data gathered from multiple sensors to monitor events inan area of interest. For example, in a disaster management setup, a largenumber of sensors can be dropped by a helicopter. Networking thesesensors can assist rescue operations by locating survivors, identifying riskyareas and making the rescue crew more aware of the overall situation.Such application of sensor networks not only can increase the efficiency ofrescue operations but also ensure the safety of the rescue crew. On themilitary side, applications of sensor networks are numerous. For example,the use of networked set of sensors can limit the need for personnelinvolvement in the usually dangerous reconnaissance missions. In addition,sensor networks can enable a more civic use of landmines by making themremotely controllable and target specific in order to prevent harmingcivilians and animals. Security applications of sensor networks includeintrusion detection and criminal hunting. However, sensor nodes areconstrained in energy supply and bandwidth. Such constraints combinedwith a typical deployment of large number of sensor nodes have posedmany challenges to the design and management of sensor networks.These challenges necessitate energy awareness at all layers of networkingprotocol stack. The issues related to physical and link layers are generallycommon for all kind of sensor applications, therefore the research on theseareas has been focused on system-level power awareness such asdynamic voltage scaling, radio communication hardware, low duty cycleissues, system partitioning, energy-aware MAC protocols. At the network
layer, the main aim is to find ways for energy-efficient route setup andreliable relaying of data from the sensor nodes to the sink so that thelifetime of the network is maximized. Routing in sensor networks is verychallenging due to several characteristics that distinguish them fromcontemporary communication and wireless ad hoc networks. First of all, itis not possible to build a global addressing scheme for the deployment ofsheer number of sensor nodes. Therefore, classical IP-based protocolscannot be applied to sensor networks. Second, in contrary to typicalcommunication networks almost all applications of sensor networks requirethe flow of sensed data from multiple regions (sources) to a particular sink.Third, generated data traffic has significant redundancy in it since multiplesensors may generate same data within the vicinity of a phenomenon.Such redundancy needs to be exploited by the routing protocols to improveenergy and bandwidth utilization. Fourth, sensor nodes are tightlyconstrained in terms of transmission power, on-board energy, processingcapacity and storage and thus require careful resource management. Dueto such differences, many new algorithms have been proposed for theproblem of routing data in sensor networks. These routing mechanismshave considered the characteristics of sensor nodes along with theapplication and architecture requirements. Data-centric protocols arequery-based and depend on the naming of desired data, which helps ineliminating many redundant transmissions. Hierarchical protocols aim atclustering the nodes so that cluster heads can do some aggregation andreduction of data in order to save energy. Location based protocols utilizethe position information to relay the data to the desired regions rather thanthe whole network. We will explore the routing mechanisms for sensornetworks developed in recent years. Each routing protocol is discussedunder the proper category. Our aim is to help better understanding of thecurrent routing protocols for wireless sensor networks.
2. Classification Of Routing ProtocolsRouting techniques are required for sending data between sensor nodesand the base stations for communication. Different routing protocols areproposed for wireless sensor network. These protocols can be classifiedaccording to different parameters.(a)Routing Protocols can be classified as Proactive, Reactive and Hybrid,based on their Mode of Functioning and Type of TargetApplications.(b)Routing protocols can be classified as Direct Communication, Flat andClustering Protocols, according to the Participation style of the Nodes.(c)Routing Protocols can be classified as Hierarchical, Data Centric andlocation based, depending on the Network Structure.2.1 Based on Mode of Functioning and Type of Target Applications2.1.1 Proactive:-In a Proactive Protocol the nodes switch on their sensors and transmitters,sense the environment and transmit the data to a BS through thepredefined route. Examples: The Low Energy Adaptive Clustering hierarchyprotocol (LEACH) utilizes this type of protocol.2.1.2 Reactive:-If there are sudden changes in the sensed attribute beyond some pre-determined threshold value, the nodes immediately react. This type ofprotocol is used in time critical applications.Examples: The Threshold sensitive Energy Efficient sensor Network(TEEN) is an example of a reactive protocol.2.1.3 Hybrid:-Hybrid protocols incorporate both proactive and reactive concepts. Theyfirst compute all routes and then improve the routes at the time of routing.Examples: Adaptive Periodic TEEN(APTEEN) is an example of a reactiveprotocol.2.2 According to the Participation style of the Nodes.2.2.1 Direct Communication:-
In this type of protocols, any node can send information to the BaseStation(BS) directly. When this is applied in a very large network, theenergy of sensor nodes may be drained quickly. Its scalability is very small.Examples: SPIN is an example of this type of protocol.2.2.2 Flat:-In this protocol, if any node needs to transmit data, it first searches for avalid route to the BS and then transmits the data. Nodes around the basestation may drain their energy quickly. Its scalability is average.Examples: Rumor Routing is an example of this type of protocol.2.2.3 Clustering Protocols:-According to the clustering protocol, the total area is divided into numbersof clusters. Each and every cluster has a cluster head (CH) and this clusterhead directly communicates with the BS. All nodes in a cluster send theirdata to their corresponding CH.Examples: TEEN is an example of this type of protocol.2.3 Depending on the Network Structure2.3.1 Data Centric:-Data centric protocols are query based and they depend on the naming ofthe desired data, thus it eliminates much redundant transmissions. The BSsends queries to a certain area for information and waits for reply from thenodes of that particular region. Since data is requested through queries,attribute based naming is required to specify the properties of the data.Depending on the query, sensors collect a particular data from the area ofinterest and this particular information is only required to transmit to the BSand thus reducing the number of transmissions.Examples: SPIN was the first data centric protocol.2.3.2 Hierarchical:-Hierarchical routing is used to perform energy efficient routing, i.e., higherenergy nodes can be used to process and send the information; low energynodes are used to perform the sensing in the area of interest.Examples: LEACH, TEEN, APTEEN.2.3.3 Location Based:-
Location based routing protocols need some location information of thesensor nodes. Location information can be obtained from GPS (GlobalPositioning System) signals, received radio signal strength, etc. Usinglocation information, an optimal path can be formed without using floodingtechniques.Examples: Geographic and Energy-Aware Routing(GEAR)
3.Data Dissemination ProtocolsData dissemination is the process by which queries or data are routed inthe sensor network. The data collected by sensor nodes has to becommunicated to the BS or to any other node interested in the data. Thenode that generates data is called a source and the information to bereported is called an event. A node which is interested in an event andseeks information about it is called a sink. Traffic Models have beendeveloped for sensor networks such as the data collection and datadissemination (diffusion) models. In the data collection model, the sourcesends the data it collects to a collection entity such as the BS. This couldbe periodic or on demand. The data is processed in the central collectionentity. Data diffusion, on the other hand, consists of a two-step process ofinterest propagation and data propagation. An interested is a descriptor fora particular, intrusion or presence of bio-agents. For every event that a sinkis interested in, it broadcasts its interest to its neighbors and periodicallyrefreshes its interest. The interest is propagated across the network andevery node maintains an interest cache of all events to be reported.3.1 FloodingIn Flooding, Each node which receives a packet broadcasts it, if themaximum hop count of the packet is not reached and node itself is not thedestination of the packet. This technique does not require complextopology maintenance or route discovery algorithms.Flooding has following disadvantages:Implosion: This is situation when duplicate messages are sent to the samenode. This occurs when a node receives copies of the same message frommany of its neighbours.Overlap: The same event may be sensed by more than one node due tooverlapping of regions of coverage. This results in their neighbors receivingduplicate reports of the same event.Resource Blindness: The flooding protocol does not consider theavailable energy at the nodes and results in many redundanttransmissions. So, it reduces the network lifetime.
3.2 GossipingGossiping is modified version of flooding, where the nodes do notbroadcast a packet, but send packets to a randomly selected neighbor.This avoids the problem of Implosion. It takes a long time for a message topropagate throughout the network. Though gossiping has considerablylower overhead than flooding, it does not guarantee that all nodes of thenetwork will receive the message. It relies on the random neighborselection to eventually propagate the message throughout the network.3.3 Rumor RoutingRumor Routing is an agent based path creation algorithm. Agents are long-lived entities created at random by nodes. These are basically packetswhich are circulated in the network to establish shortest path to events thatthey encounter. They can also perform path optimizations at nodes theyvisit. When agent finds a node whose path to an event is longer than itsown, it updates the nodes routing table.Figure 3.1 illustrates the working of Rumor Routing algorithm. In figure3.1(a), the agent has initially recorded a path distance 2 to event E1. NodeAs table shows that it is at a distance 3 from event E1 and distance 2 fromE2. When the agent visits node A, i+t updates its own path stateinformation to include the path to event E2. The updating is with one hopgreater distance than what it found in A, to account for the hop between
any neighbor of A that the agent will visit next, and A. It also optimizes thepath to e1 recorded at node A to the shorter path through node B. Theupdated status of the agent and node table is shown in figure 3.1(b). When a query is generated at a sink, it is sent on a random walk withthe hope that it will find a path leading to the required event. This is basedon high probability of two straight lines intersecting on a planar graph,assuming the network topology is like a planar graph, and the pathsestablished can be approximated by straight lines owing to high density ofthe nodes. If a query does not find an event path, the sink times out anduses flooding as last resort to propagate the query. For instance, as in figure 3.1(c), suppose a query for event E1 isgenerated by node P. Through a random walk, it reaches A, where it findsthe previously established path to E1. Hence, the query is directed to E1through node B, as indicated by As table.3.4 Sequential Assignment Routing :-The Sequential Assignment Routing(SAR) creates multiple trees ,where theroot of each tree is a one hop neighbor of sink. Each tree grows outwardfrom the sink and avoids nodes with low throughput or high delay. At theend of the procedure, most nodes belong to multiple trees. An instance oftree formation is illustrated in figure.
The tree rooted at A and B. Two of the one hop neighbors of the sink areshown. Node C belongs to both trees and has path length of 3 and 5respectively to the sink, using the two trees. Each sensor node records twoparameters about each path through it:1. The available energy resources on the path.2. An additive quality of service(QoS) metric such as delay.This allows a node to choose one path from among many to relay itsmessage to the sink. The SAR chooses a path with the high estimatedenergy resources and provisions can be made to accommodate packets ofdifferent properties. A weighted QoS metric is used to handle prioritizedpackets which computed as a product of priority level and delay. Therouting ensures that the same weighted QoS metric is maintained. Thus,higher priority packets take lower delay paths and lower priority packetshave to use the paths of greater delay. e.g. If node C generates a packet ofpriority 3, it follows the longer path along tree B, and a packet of priority 5
will follow the shorter path along tree A. So that the priority X delay QoSmetric is maintained. SAR minimizes the average weighted QoS metricover the lifetime of the network. The sink periodically triggers a metricupdate to reflect the changes in available energy resources after sometransmissions.3.5 Direct DiffusionThis protocol is useful in scenario where the sensor nodes themselvesgenerate requests/queries for data sensed by other nodes, instead of allqueries arising only from a BS. Hence the sink for the query could be a BSor a sensor node. The direct diffusion routing protocol improves on datadiffusion using interest gradients. Each sensor node names its data withone or more attributes and other nodes express their interest depending onthese attributes. Attribute value pairs can be used to describe an interest inintrusion data as follows. The sink has to periodically refresh its interest if itstill requires the data to be reported it. Data is propagated along thereverse path of the interest propagation. Each path is associated with agradient that is formed at the time of interest propagation. Each path isassociated with the gradient that is formed at the time of interestpropagation. While the positive gradients encourage the data flow along thepath, Negative gradients inhibit the distribution of data along a particularpath. The strength of the interest is different toward different neighbors,resulting into source to sink paths with different gradients. The gradientcorresponding to an interest is derived from the interval/data-rate fieldspecified in the interest.This model uses data naming by attributes and local data transformation toreflect the data centric nature of sensor network operations. The localoperations of Data aggregation are application-specific gradient model. Thenetwork wide results of local interaction by regulating the flow of data alongdifferent paths depending on the expressed interest.3.6 Sensor Protocol for Information via Negotiation(SPIN)SPIN uses negotiation and resources and adaption to address thedeficiencies of flooding. Negotiation reduces overlap and implosion, and athreshold based resource-aware operation is used to prolong networklifetime. Meta-data, or data describing data, is transmitted instead of rowdata. This requires fewer bytes and can be in an application-specificformat.
SPIN has three types of messages: ADV, REQ, and DATA. A sensor nodebroadcasts an ADV containing meta-data describing actual data. If aneighbor is interested in the data, it sends REQ for the data. Then thesensor node sends the actual DATA to the neighbor. The neighbor againsends ADVs to its neighbors and this process continues to disseminate thedata throughout the network. the simple version is shown in figure.SPIN is based on data-centric routing, where the nodes advertise theavailable data through an ADV and wait for requests from interested nodes.SPIN-2 expands on SPIN, using an energy or resource threshold to reduceparticipation. A node may participate in the ADV-REQ-DATA handshakeonly if it has sufficient resources above a threshold.3.7 Geographic Hash TableGeographic Hash Table is a system based on data centric storage inspiredby internet scale distributed hash table systems such as chard andTapestry, GHT hashes keys into geographic co-ordinates and stores a pairat the sensor node nearest to the hash value. The calculated hash value ismapped onto a unique node consistently, so that queries for the data canbe routed to the correct node. Stored data is replicated to ensureredundancy in case of node failures and consistently protocol is used tomaintain the replicated data. The data is distributed among nodes such thatit is scalable and the storage load is balanced. GHT is more effective in
large network where a large number of events are detected but not all arequeried. In this case data observed is stored in a distributed manner acrossall nodes, instead of being routed to central external storage. Queries arerouted to the nearest node which contains a copy of the relevant data. Thismakes the storage and traffic distribution uniform.
4.Data Gathering ProtocolsThe objective of the data-gathering problem is to transmit the sensed datafrom each sensor node to a BS. One round is defined as the BS collectingdata from all the sensornodes once. The goal of algorithms which implement data gathering is tomaximize the number of rounds of communication before the nodes dieand the network becomes inoperable. This means minimum energy shouldbe consumed and the transmission should occur with minimum delays,which are conflicting requirements. Hence, the energy X delay metric isused to compare algorithms, since this metric measures speedy andenergy efficient data gathering. A few algorithms that implement datagathering are discussed below.4.1 Direct TransmissionAll sensor nodes transmit their data directly to BS. This is extremelyexpensive in terms of energy consumed, since the BS may be very faraway from some nodes. Also, nodes must take turns while transmitting tothe BS to avoid collision , so the media access delay is also large. Hence,this scheme performs poorly with respect to the energy X delay matrix.4.2 Power Efficient Gathering for Sensor Information SystemsPower Efficient Gathering for Sensor Information Systems (PEGASIS) is adata-gathering protocol based on the assumption that all sensor nodesknow the location of every other node, that is, the topology information isavailable to all nodes. Also, any node has the required transmission rangeto reach the BS in one-hop, when it is select as a leader. The goals ofPEGASIS are as follows: Minimize the distance over which each node transmits. Minimize the broadcasting overhead. Minimize the number of messages that need to be sent to the BS. Distribute the energy consumption equally across all nodes.
A greedy algorithm is used to construct a chain of sensor nodes, startingfrom the node farthest from the BS. At each step, the nearest neighborwhich has not been visited is added to the chain. The chain is constructeda priory, before data transmission begins and is reconstructed when nodesdie out. At every node, data diffusion is carried out. So, that only onemessage is passed on from one node to next. A node which is designatedas the leader finally transmits one message to BS.Leadership is transferred in sequential order and a token is passed. So thatthe nodes know in which direction to pass messages in order to reach theleader. A possible chain formation is illustrated in figure. The delay involvedin message reaching the BS is O(N), where N is the total number of nodesin the network.4.3. LEACHLow-energy adaptive clustering hierarchy is one of the most popularhierarchical routing algorithms for sensor networks. The idea is to formclusters of the sensor nodes based on the received signal strength and uselocal cluster heads as routers to the sink. This will save energy since thetransmissions will only be done by such cluster heads rather than all sensornodes. Optimal number of cluster heads is estimated to be 5% of the totalnumber of nodes.All the data processing such as data fusion and aggregation are local to thecluster. Cluster heads change randomly over time in order to balance the
energy dissipation of nodes. This decision is made by the node choosing arandom number between 0 and 1. The node becomes a cluster head forthe current round if the number is less than the following threshold:where p is the desired percentage of cluster heads (e.g. 0.05), r is thecurrent round, and G is the set of nodes that have not been cluster headsin the last 1=p rounds. LEACH achieves over a factor of 7 reduction inenergy dissipation compared to direct communication and a factor of 4–8compared to the minimum transmission energy routing protocol. The nodesdie randomly and dynamic clustering increases lifetime of the system.LEACH is completely distributed and requires no global knowledge ofnetwork. However, LEACH uses single-hop routing where each node cantransmit directly to the cluster-head and the sink. Therefore, it is notapplicable to networks deployed in large regions. Furthermore, the idea ofdynamic clustering brings extra overhead, e.g. head changes,advertisements etc., which may diminish the gain in energy consumption.
5.References :-1. Wireless sensor networks: a survey I.F. Akyildiz, W. Su*, Y. Sankarasubramaniam, E. Cayirci2. Ad Hoc Wireless Networks By,C.Shiva Ram Murthy and B.S.Manoj3. A survey on routing protocols for wireless sensor networks Kemal Akkaya *, Mohamed Younis4.www.wikipedia.org/wiki/Wireless sensor network