In this paper, we propose to design an energy efficientreliable multipath routing algorithm for data gathering inwireless sensor networks. This protocol is intended to providea reliable transmission environment with low energyconsumption, by efficiently utilizing the energy availability of the forwarding nodes to gather and distribute the data to sink,according to its requirements.II.
Deepak Ganesan et al  have proposed a highly resilient;energy-efficient multipath routing protocol. It mainlydiscusses energy efficient recovery from failures, bydiscovering alternate paths. But it fails to consider the QoSqualities of the routes, when constructing multiple paths.R Vidhyapriya et al , have developed an adaptivemultipath routing protocol which spreads the traffic over thenodes lying on different possible paths between the source andthe sink, in proportion to their residual energy and receivedsignal strength. But it transmits all the packets across the multipaths, without considering the category of data.Weifa Liang et al , have proposed a maximum network lifetime routing (MNL) algorithm to maximize the network lifetime while gathering online queries. In this protocol, thesink constructs a tree towards the source node, based on theresidual energies of the nodes. But it does not consider thereliability of the transmitted data, since there will be largevolumes of data involved, when there are continuous queriesat the sink.Ye Ming Lu et al , have proposed an energy efficientmultipath routing algorithm which uses a load balancingalgorithm to distribute the traffic over multiple disjoint paths.For energy efficiency, it uses the residual energy in the link cost function. But it does not consider aggregating similar dataalong multiple paths.Yuzhe Liu et al , have proposed a priority basedmultipath routing protocol. It forwards the disseminated databased on the priority information accumulated hop count orremaining power resource. It uses either shortest path orenergy-efficient path based on the priority tag. But thisprotocol does not consider the category of data and nature of queries.Antoine B. Bagula et al , have proposed an energyconstrained multipath routing. It minimizes the number of paths used in forwarding the data to the sink, there byminimizing the energy. But it does not discuss the sink’sinterest and reliability of data.Octav Chipara et al , have proposed a real-time poweraware routing (RPAR) protocol. It addresses importantpractical issues in wireless sensor networks, including lossylinks, scalability, and severe memory and bandwidthconstraints.III.
A. Protocol Overview
In this paper, we propose an energy efficient reliablemultipath routing algorithm for data gathering in wirelesssensor networks. It consists of three phases:1.
Periodic Interest Propagation by the sink 2.
Energy Efficient Multipah Tree Construction3.
Packet DispersionIn the first phase, the sink periodically broadcast aninterest message containing its required data model, to all thenodes. In the second phase, we construct a multi path tree, inwhich nodes are selected based on their residual energy level.In the third phase, data sources of similar interests aregathered and transmitted towards the sink across the energy-efficient tree. When data sources cannot be aggregated, theyare dispersed along multiple paths using erasure codingtechnique .
B. Periodic Interest Propagation by the Sink
A sink generates an interest message that identifies itsrequirement in wireless sensor networks which is thenpropagated throughout the network. On receiving an interestmessage, the source transmits the corresponding data. The datapackets having similar interests are collected and aggregated atintermediate aggregators. The sink does not have anyinformation on the availability of data while transmitting thefirst interest message. So the sink simply broadcasts interestmessage to all its neighbors. Interest message contains theInterest Id, Description and Timestamp. The features of shortest path algorithm can be used for interest messagepropagation.An interest table is maintained by each node whichcontains the fields Interest Id, Sender Id, Cost of the messagein terms of hop count and Timestamp. On receiving an interestmessage the node will look up in its interest for the receivedinterest message. An interest table makes only one entry perdata type from a particular sink.When a node delivers the first interest message, it is addedin the interest table with its parameters. The interest messageis then rebroadcast to other nodes. It checks the interest table,if a duplicate interest message is received by a node. Theduplicate message is dropped when the cost of it is higher thanthe cost of the earlier message; else it is updated in the tableand then forwarded to the next node.The proposed protocol consists of a periodic interestpropagation phase. Since the interest is a soft state, it is veryoften refreshed by the sink. Refreshing is essential since it isimpossible to transmit interest reliably across the network andthe refresh rate is a protocol design parameter. To propagatethe interest based on the previously cached data, eitherflooding or directional propagation may be used.
C. Energy Efficient Multipath Tree Construction
We propose a heuristic algorithm for the tree construction.We consider the wireless sensor network M as a directedgraph G (N, E). Let the set of nodes N consisting of sensors
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 2, May 201060http://sites.google.com/site/ijcsis/ISSN 1947-5500