sdrina: secured and reliable data routing approach …

1

Int. J. Elec&Electr.Eng&Telecoms. 2014 Ajey Kumar R and Keerthi D S, 2014

SDRINA: SECURED AND RELIABLE DATAROUTING APPROACH FOR IN-NETWORKAGGREGATION IN WIRELESS SENSOR

NETWORKS

Ajey Kumar R1* and Keerthi D S1

*Corresponding Author: Ajey Kumar R,[email protected]

The fundamental challenge in the design of Wireless Sensor Networks (WSNs) is to maximizetheir lifetime. Data aggregation has emerged as a basic approach in WSNs in order to reducethe number of transmissions of sensor nodes, hence minimizing the overall power consumptionin the network. Data aggregation is affected by several factors, such as the placement ofaggregation points, the aggregation function, and the density of sensors in the network. Thedetermination of an optimal selection of aggregation points is thus extremely important. Wepresent exact and approximate algorithms to find the minimum number of aggregation points inorder to maximize the network lifetime. Our results clearly indicate that the routing tree built bySDRINA provides the best aggregation quality when compared to existing algorithms. The obtainedresults show that our proposed solution outperforms in different performance metrics and indifferent scenarios required by WSNs. We also study the tradeoffs between energy savingsand the potential delay involved in the data aggregation process.

Keywords: In-network aggregation, Routing protocol, Wireless sensor networks, Clusters,Hop trees

INTRODUCTIONInformation gathering is a fast growing andchallenging field in today's world of computing.Sensors provide a cheap and easy solution tothese applications especially in theinhospitable and low-maintenance areaswhere conventional approaches prove to be

ISSN 2319 – 2518 www.ijeetc.comVol. 3, No. 3, July 2014

© 2014 IJEETC. All Rights Reserved

Int. J. Elec&Electr.Eng&Telecoms. 2014

1 Department of Electronics and Communication Engineering, Malnad Collage of Engineering, Hassan, Karnataka, India.

very costly. Sensors are tiny devices that arecapable of gathering physical information likeheat, light or motion of an object orenvironment. Generally, a sensor node doesnot have sufficient power to send the data ormessage directly to the base station. Hence,along with sensing the data the sensor node

Research Paper

2


act as a router to propagate the data of itsneighbour. In wireless sensor network, due tothe high density of nodes, the redundant datawill be detected by neighbouring nodes whilesensing an event. In order to save energy allthese redundant data will be aggregated at anintermediate node and then it will send to thesink node (Zhang et al., 2011).

Wireless sensor network has some keyconstraints such as limited energy resources,lack of infrastructure. A possible strategy tooptimize the routing task is to use the availableprocessing capacity provided by theintermediate sensor nodes along the routingpaths. This is known as data-centric routing orin-network data aggregation (Abdel Salamand Olariu, 2009; and Villas et al., 2010). Oneof the main challenges in routing algorithmsfor WSNs is how to guarantee the delivery ofthe sensed data even in the presence of nodesfailures and interruptions in communications.These failures become even more criticalwhen data aggregation is performed along therouting paths since packets with aggregateddata contain information from various sourcesand whenever one of these packets is lost aconsiderable amount of information will alsobe lost. In the context of WSN, dataaggregation aware routing protocols shouldpresent some desirable characteristics suchas: a reduced number of messages for settingup a routing tree, maximized number ofoverlapping routes, high aggregation rate andalso a reliable data transmission. In order toovercome these challenges, in this work, wepropose a novel Data Routing algorithm forIn-Network Aggregation for WSNs, which werefer to as SDRINA algorithm (Chatzigiannakiset al., 2005; and Villas et al., 2010).

RELATED WORKRouting Challenges and DesignIssues in WSN• Node deployment

• Energy consumption without losingaccuracy

• Node addressing and location awareness

• Sensor network are mostly data centric

• Reliable data aggregation and transmission

• Node/link heterogeneity and networkdynamics

• Transmission media

• Coverage area

• Scalability

• QoS

Hierarchical Routing ProtocolTraditional routing protocols for WSN may notbe optimal in terms of energy consumption.Clustering techniques can be efficient in termsof energy and scalability (Jamal Al-Karaki andAhmad Kamal, 2004). The objective ofclustering is to minimize the total transmissionpower aggregated over the nodes. Everycluster selects a Cluster Head (CH)responsible for coordinating the datatransmission among the nodes in a cluster,which collects the data and transmit it to theBase Station (BS) (Figure 1).

Main advantages of hierarchical routingprotocols are:

• Minimization of intra-cluster and inter clusterenergy consumption.

• Scalability and Prolong network life time.

3


• Reducing data packet delay.

• Handling node heterogeneity.

Existing System: InformationFusion-Based Role Assignment(InFRA)The Information Fusion-based RoleAssignment (In-FRA) algorithm builds a clusterfor each event including only those nodes thatwere able to detect it. Then, cluster headsmerge the data within the cluster and send theresult toward the sink node. The InFRAalgorithm aims at building the shortest pathtree that maximizes information fusion (Figure2). A disadvantage of the InFRA algorithm isthat for each new event that arises in thenetwork, the information about the event mustbe flooded throughout the network to informother nodes about its occurrence and toupdate the aggregated coordinators distance.This procedure increases the communicationcost of the algorithm and thus limits itsscalability (Nakamura et al., 2006). In thecontext of WSN, data aggregation awarerouting protocols should present some

desirable characteristics such as: a reducednumber of messages for setting up a routingtree, maximized number of overlapping routes,high aggregation rate, reliable datatransmission and also a route repairmechanism. In order to overcome thesechallenges, in this work, we propose a Securedand Reliable Data Routing algorithm forNetwork Aggregation in WSNs, which we referto as SDRINA algorithm. Our proposedalgorithm was conceived to maximizeinformation fusion along the communicationroute in reliable way, through a fault-tolerantrouting mechanism.

Proposed System: SDRINA(Enhancement)The proposed system presents a secure andreliable data gathering scheme, which ensuresthat the trajectory of the mobile anchor nodeminimizes the energy consumption of nodesand guarantees that all of the sensor nodescan retain higher energy levels, a single mobileanchor node moves randomly through thesensing field broadcasting periodic three

Figure 1: Clustering of Sensor Network Figure 2: NAM Output of Existing System(InFRA)

4


beacon messages containing its currentcoordinates. The locations of the individualsensor nodes are determined by exploiting thefact that the perpendicular bisector of a chordof a mobile anchor passes through the centerof the circle twice. The mobile anchor collectsthe data from sensor nodes (CHs) and deliversthese aggregated data to the sink node throughintermediate nodes (Figure 3).

SDRINA: SECURED ANDRELIABLE DATA ROUTINGFOR IN-NETWORKAGGREGATION IN WSNSThe main goal of DRINA is to build a routingtree and find out the shortest path whichconnects all source nodes to the sink, whilemaximizing the data aggregations. In SDRINAfollowing roles are consider for building therouting tree (Leandro Villas et al., 2013).

Collaborator: It is the node which detects anevent and reports the collected data to theCoordinator node.

Coordinator: It is the node that also detectan event but after using election algorithm. Thisnode is responsible for aggregating thecollected data received from other collaboratorand send aggregated results to the sink node.

Sink Node: This node is interested inreceiving the data from set of coordinatornodes and collaborator node.

Relay Node: This is the intermediate nodebetween Coordinator and Sink node and it isresponsible for forwarding data toward theSink.

Mobile Anchor: This node is responsible forcoordinating the location of each Clustermembers and it collects the data fromCoordinator (CHs) and delivers theseaggregated data to the Sink node throughintermediate nodes.

SDRINA algorithm is divided into fivephase:

1. Constructing the Hop tree from sensornodes to the Sink node.

2. Cluster formation and electing a clusterhead among the Collaborator whichbecomes a Coordinator.

3. Setting up a route for reliable delivery ofdata packets and updating the hop tree.

4. Secure and Reliable Data gatheringscheme.

5. Route Repair Mechanism.

Constructing the Hop TreeHop tree is constructed from sensor node tothe Sink node. The distance is computed inhops from the sink node to each node in thenetwork. The Sink node sends HopConfiguration Message (HCM) to all nodes

Figure 3: NAM Output of Proposed System(SDRINA)

5


through flooding. This is the distance throughwhich the HCM message has passed calledHopToTree. The initial value of HopToTree is1 at the sink. This HCM message is passedto its neighbours, when receiving this message,each node verifies if the stored HopToTreevalue is greater than the value of HopToTreein the received HCM message. If the conditionis true then the node updates the storedHopToTree value with the value in the HCMmessage else the node discards the receivedHCM message (Figure 4).

Cluster FormationFormation of cluster and cluster head electionalgorithm should be done when one or morenodes detect the same event. All sensingnodes are eligible for this election. ClusterHead will be one that is closest to the Sinknode or node that is closest to an alreadyestablished route. In case of a tie, i.e., two ormore concurrent nodes have the samedistance in hops to the sink (or to anestablished route) then Energy level is used

as a tiebreak criterion. Based on the result ofthis algorithm, roles are assigned to the nodes.Elected leader will be declared as theCoordinator and remaining nodes that detectthe same event will be the Collaborators. Theadvantage of this algorithm is that theinformation collected by the nodes which aresensing the same event will be aggregated atthe single point (called aggregation point)which is more efficient.

Routing Formation and Hop TreeUpdatesNew route is established by the Coordinatorfor the event dissemination. To establish theroute, Coordinator sends a routeestablishment message to its NextHop node.When this node receives a route establishmentmessage, it establishes the route and forwardsthe message to its NextHop, after the routeestablishment, HopToTree updating processis started. The main aim of this phase is toupdate the value of the HopToTree in all nodesso that newly established route can taken intoconsideration. Relay nodes are responsiblefor this process. Each node will send only onepacket so that the whole cost of this processis same as flooding.

Secured and Reliable DataGathering SchemeMobile anchor is responsible for coordinatingthe location of each Cluster members. Itminimizes the energy consumption of nodesand guarantees that all of the sensor nodescan retain higher energy levels, a single mobileanchor node moves randomly through thesensing field broadcasting periodic threebeacon messages containing its currentcoordinates and it collects the data from

Figure 4: SDRINA Routing TreeEstablishment and Updating

6


Coordinator and delivers to Sink node. Thecloser the event take place, the lower thecommunication cost, thus the best case willbe achieved when the events happen on therouting tree, and mobile anchor is responsiblefor this action.

Route Repair MechanismA route repair mechanism is used to sendinformation in a reliable way. SDRINAalgorithm has piggybacked and ACK-basedroute repair mechanism, which consists of twoparts: failure detection at the NextHop node,and selection of a new NextHop. When a relay

node needs to forward data to its NextHopnode, it simply sends the data packet, sets atimeout, and waits for the retransmission of thedata packet by its NextHop. This re-transmission is also considered an ACKmessage. If the sender receives its ACK fromthe NextHop node, it can infer that the NextHopnode is alive and, for now, everything is ok.However, if the sender node does not receivethe ACK from the NextHop node within thepredetermined timeout, then it recognizes thatparticular node get failed and another oneshould be selected as the new NextHop node.For this, the sender chooses the neighbor withthe lowest hop-to-tree level to be its newNextHop; in case of a tie, it chooses theneighbor with the highest energy level. A newlyreconstructed path is created after therepairing mechanism is used forretransmission, as shown in Figures 5a-5b.

RESULTS AND ANALYSISThe simulations were performed usingNetwork Simulator-2 (Ns-2), particularlypopular in the wireless networking community.The traffic sources are Continuous Bit Rate(CBR). The source-destination pairs arespread randomly over the network. Themobility model uses ‘random waypoint model’

Parameter Value

Simulator Network Simulator-2

Simulation area 700 x 700

No. of nodes 30

Communication radius (m) 50

Initial energy (J) 100

Traffic type CBR (UDP)

Node placement model Random waypoint

Table 1: Simulation Parameters

Figure 5: Example of Path Repair

(a) Region with Destroyed Nodes

(b) Repaired Path

7


in a rectangular filed of 700 m x 700 m with20-30 nodes.

This section presents the performanceevaluation for SDRINA. Since SDRINA is animprovement on InFRA, we compare boththese existing and proposed algorithm usingthe following performance metrics:

ThroughputThis metric indicates that the packets perprocessed data, Figure 6, shows a graph forthe InFRA and SDRINA algorithm on the basisof throughput as a function of pause time andusing different number of traffic sources. It isthe rate between the total packets transmitted(data and control packets) and the number ofdata packet received by the sink. It is clearedthat SDRINA out performs than InFRA. This isbecause SDRINA has the route repairmechanism and due to that number ofsuccessful and reliable data transmission rateincreases.

Packet Delivery Rate (PDR)This metric indicates that the rate at whichnumber of packets that reaches the sink node,

in the SDRINA algorithm data packettransmission increases when the aggregationrate increases in the built tree, SDRINA has abetter aggregation rate than InFRA this is due tothe fact that lost packets with aggregated dataare retransmitted. On the other hand, in InFRA itdoes not retransmit the data packets which arelost due to communication failures (Figure 7).

Figure 6: Throughput

Figure 7: Packet Delivery Rate (PDR)

End To End DelayThis metric indicates the quality of the routingtree built by the algorithms, the delay time

Figure 8: End to End Delay

8


decrease when the probability of thecommunication failure decreases. For bestcase scenario i.e., when probability ofcommunication failure is less InFRA has lessdelay time. For worst case scenario i.e., whenprobability of communication failure is highthen SDRINA exhibits the less delay time andhas better compared to InFRA this is due toSDRINA has route repair mechanism whichleads less communication failure (Figure 8).

Energy ConsumptionEnergy consumption is the total energyconsumed by the total number of edges in therouting tree structure built by the algorithm.Energy consumption in SDRINA is less this isbecause it needs less control messages tobuild the routing tree when compared toInFRA, in SDRINA fault tolerant and In-networking aggregation can often be used todecrease the communication cost byeliminating redundancy and forwarding onlysmaller aggregated information. Sinceminimal communication load and lessprobability of communication failure leads

directly to energy savings, which extends thenetwork lifetime (Figure 9).

An aggregation aware routing protocolplays an important role in event-based WSNs.The goal of this performance evaluation is acomparison between InFRA and SDRINA data

Figure 9: Energy Consumption

Table 2: Performance Comparison

ParametersExisting

System (InFRA)Proposed

System (SDRINA)

RoutingCategory

Aggregationnode

Scalability

Route Repairmechanism

Facts

Findings

Cluster based

Cluster headsand Relaynodes

Low

No

Maximizeoverlap routes

Low scalabilityand highoverhead

Tree-basedCluster

Cluster heads,Mobile anchorand Relay nodes

High

Yes

Reliable routingand Minimizationof delay andenergyconsumption

Less efficient fordynamic routes

aggregation routing protocols. SDRINA in oursimulation experiment shows to have theoverall best performance, SDRINA performsbetter at high packet delivery rate, has a highthroughput and better performs in both energysavings, routing delay as compared to InFRA.

CONCLUSIONIn this work we present the securedaggregation aware routing protocol SDRINAin order to achieve two main goals secured,reliable data transmission and increase theenergy savings. By maximizing theaggregation points and offering a fault tolerantmechanism can eliminate the redundanttransmission which leads reduced

9


communication load. The obtained resultsclearly show that SDRINA outperformed theexisting data aggregation algorithms fordifferent metrics.

The proposed algorithm has some keyaspects required by WSNs aggregationaware routing algorithms such as a reducednumber of messages for setting up a routingtree, maximized number of overlapping routes,high aggregation rate, secured and reliabledata aggregation and transmission. As futurework, DRINA performs efficient only in the caseof static events of fixed radius. Our future workmay consider the Dynamic sizes of events.New strategies will be devised to control thewaiting time for aggregator nodes based ontwo criteria: average distance of the eventcoordinators, spatial and semantics eventcorrelation.

ACKNOWLEDGMENTThe authors wish to thank the reviewers andeditors for their valuable suggestions andexpert comments that help improve the paper.

REFERENCES1. AbdelSalam H S and Olariu S (2009), “A

Lightweight Skeleton ConstructionAlgorithm for Self-Organizing SensorNetworks”, Proc. IEEE Int’l Conf. Comm.(ICC), p. 5.

2. Al-Karaki J and Kamal A (2004), “RoutingTechniques in Wireless Sensor Networks:A Survey”, IEEE Wireless Comm. ,Vol. 11, No. 6, pp. 6-28.

3. Al-Karaki J, Ul-Mustafa R and Kamal A(2004), “Data Aggregation in-WirelessSensor Networks-Exact and ApproximateAlgorithms”, Proc. High Performance

Switching and Routing Workshop(HPSR’04), pp. 241-245.

4. Akyildiz I F, Su W, SankarasubramaniamY and Cyirci E (2002), “Wireless SensorNetworks: A Survey”, Computer Networks,Vol. 38, No. 4, pp. 393-422.

5. Anastasi G, Conti M, Francesco M andPassarella A (2009), “EnergyConservation in Wireless SensorNetworks: A Survey”, Ad Hoc Networks,Vol. 7, No. 3, pp. 537-568.

6. Boukerche A, Araujo R B and Villas L(2007), “Optimal Route Selection forHighly Dynamic Wireless Sensor andActor Network Environment”, Proc. 10th

ACM Symp. Modeling, Analysis,andSimulation of Wireless and MobileSystems (MSWiM’07), pp. 21-27.

7. Boukerche A, Turgut B, Aydin N, AhmadM Z, Boloni L and Turgut D (2011), “SurveyPaper: Routing Protocols in Ad HocNetworks: A Survey”, IEEE ComputerNetworks, Vol. 55, pp. 3032-3080.

8. Chatzigiannakis I, Nikoletseas S andSpirakis P G (2005), “Efficient and RobustProtocols for Local Detection andPropagation in Smart Dust Networks”,Mobile Networks and Applications,Vol. 10, Nos. 1/2, pp. 133-149.

9. Fasolo E, Rossi M, Widmer J and ZorziM (2007), “In-Network AggregationTechniques for Wireless SensorNetworks: A Survey”, IEEE WirelessComm., Vol. 14, No. 2, pp. 70-87.

10. Hu F, Cao X and May C (2005),“Optimized Scheduling for DataAggregation in Wireless Sensor

10


Networks”, Proc. Int’l Conf. InformationTechnology: Coding and Computing(ITCC’05), pp. 557- 561.

11. Intanagonwiwat C, Govindan R, Estrin D,Heidemann J and Silva F (2003),“Directed Diffusion for Wireless SensorNetworking”, IEEE/ACM Trans.Networking, Vol. 11, No. 1, pp. 2-16.

12. Jamal N Al-Karaki and Ahmad E Kamal(2004), “Routing Techniques in WirelessSensor Networks: A Survey”, IEEEWireless Communications, December.

13. Krishnamachari B, Estrin D and WickerS B (2002), “The Impact of DataAggregation in Wireless SensorNetworks”, Proc. 22nd Int’l Conf.Distributed Computing Systems(ICDCSW’02), pp. 575- 578.

14. Leandro Villas, Azzedine Boukerche,Heitor S Ramos, Horacio A B F deOliveira, Regina B de Araujo and AntonioA F Loureiro (2013), “DRINA: ALightweight and Reliable RoutingApproach for In-Network Aggregation inWireless Sensor Networks”, IEEEComputer Networks, Vol. 62, No. 4,pp. 676-689.

15. Nakamura E F, de Oliveira H A B F,Pontello L F and Loureiro A A F (2006),“On Demand Role Assignment for Event-Detection in Sensor Networks”, Proc.IEEE 11th Symp. Computers and Comm.(ISCC’06), pp. 941-947.

16. Nakamura E F, Loureiro A A F and FreryA C (2007), “Information Fusion forWireless Sensor Networks: Methods,Models, and Classifications”, ACMComputing Surveys, Vol. 39, No. 3,pp. 9-1/9-55.

17. Solis I and Obraczka K (2004), “TheImpact of Timing in Data Aggregation forSensor Networks”, IEEE Int’l Conf.Comm., Vol. 6, pp. 3640-3645.

18. Villas L, Boukerche A, de Araujo R Band Loureiro A A F (2010), “HighlyDynamic Routing Protocol for DataAggregation in Sensor Networks”, Proc.IEEE Symp. Computers and Comm(ISCC), pp. 496-502.

19. Zhang H, Li L, Yan X and Li X (2011), “ALoad Balancing Clustering Algorithm ofWSN for Data Gathering”, 978-1-4577-0536-6/11, IEEE.

sdrina: secured and reliable data routing approach …

Documents