Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804798 | P a g ePerformance Analysis and Enhanced Collection Tree Protocol inMobile EnvironmentBijal Chawla*, Dhaval Patel**, Chandresh Parekh***, MirenKaramta****, Dr. M B Potdar******(Department of Wireless Mobile Computing, Gujarat Technological University, Gujarat, India)** (Department of Wireless Mobile Computing, Gujarat Technological University, Gujarat, India)*** (Head of Department in Electronics and Communication, Government Engineering College, Gujarat, India)**** (Project Scientist at Bhaskaracharya Institute for Space Applications and Geo-Informatics, Gujarat, India)***** (Project Director at Bhaskaracharya Institute for Space Applications and Geo-Informatics, Gujarat, India)ABSTRACTWireless Sensor Network (WSN) isgrowing over widely and with its growth moreemphasis has been seen in mobile sensor nodes.Collection tree protocol (CTP) is one of the mostpromising data collection tree protocol. CTP isvery efficient routing protocol and it forms thebasis for other protocols. CTP has a wide rangeof applications in static WSN. In this paper wewill analyze the performance of CTP in staticand mobile environment. We will also identifyimportant factors that contribute to thedegradation of CTP’s performance in mobileenvironment. The paper will also show thesolution based on the analyses which results inincreased data delivery ratio without any controloverhead.Keywords-Collection Tree Protocol (CTP), DataCollection, Routing Protocol, Wireless SensorNetwork (WSN)1. IntroductionSince a long time in history of WirelessSensor Network (WSN) there is a large number ofdata collection protocols that has been used. Amongthese, the Collection Tree Protocol (CTP) is widelyregarded as the reference protocol for data collection. Currently TinyOS is one of the operating systemwhich supports an implementation of CTP.WSN has been growing rapidly and with itsincreasing growth mobile wireless sensor networksare becoming common. In mobile sensor networksthere are mobile nodes which are moving. Due tothese moving nodes network topology keeps onchanging. Thus there must be routing protocol forsuch network that can transmit the data correctly andreliably to the sink node. As described previouslyCTP is one of the most promising data collection treeprotocol. CTP is being very efficient in staticenvironment and has large number of applicationsalso. In this paper we will show that in mobileenvironment the performance of CTP is not as mucheffectual as compared to static environment.We will also identify the important factorswhich are affecting CTP and are responsible fordegradation of CTP in mobile environment.2. CTP ModulesAs shown in the Fig. 1  there are threemain logical software components of CTP: RoutingEngine (RE), Forwarding Engine (FE), and LinkEstimator (LE).2.1. Routing Engine (RE)Routing Engine is basically concerned withthe sending and receiving of beacons. It also takescare of creating and updating routing table. Routingtable consists of information related to the neighborswhich comes helpful during parent selection inrouting mechanism? This table is being updated onreceiving beacons at a fixed interval. Apart fromthese routing table also consists of a metric whichshows the quality of a link.Fig. 1. Module Interaction and Message Flow in CTPIn CTP this metric is called ETX (ExpectedTransmission). A node having an ETX equal to n isexpected to be able to deliver a data packet to thesink with a total of n transmissions, on average .
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804799 | P a g eThis ETX value is being exchangedbetween the nodes along with the beacons. Whileselecting a parent the node will compare its neighboron the basis of ETX value and will select a node withthe lowest ETX value as its parent.2.2. Forwarding Engine (FE)Forwarding Engine generally forwards datapackets either from the application layer or from theneighbors. The FE is also responsible of detectingand repairing routing loops as well as suppressingduplicate packets .2.3. Link Estimator (LE)Link Estimator determines the inbound andoutbound link quality of the neighboring nodes for 1-hop communication links. The LE computes the 1-hop ETX by collecting statistics over the number ofbeacons received and the number of successfullytransmitted data packets .Inbound metric is calculated as a ratiobetween the total numbers of beacon sent by aneighbor over the fraction of received beacon. Whilethe Outbound metric is calculated as the number oftransmission attempts required by a node tosuccessfully deliver a data packet to its neighbor.3. Key Challenges of CTP3.1. Packet DuplicatesDuplication of packets occurs when a nodereceives a data packet, sends an acknowledgmentpacket but the acknowledgment is not received. Dueto this the sender will send the data packet again andthe receiver will receive it twice. This duplicationwill effect over number of hopes as duplication isexponential.3.2. Routing LoopsNormally in parent selection a node willselect a node with lower ETX value but if it selectsan ETX value higher the previous one then the loopoccurs. If the new route being selected contains anode which was a descendant earlier, then a loopoccurs.3.3. Link DynamicsPeriodic beaconing is being used in aprotocol to maintain the topology and estimate linkqualities. An efficient link quality estimationtechnique is vital for the performance of a collectionprotocol. The beaconing rate introduces a tradeoffbetween agility and efficiency: a faster rate leads toa more agile network but higher cost, while a lowerrate leads to a slower-to-adapt network and lowercost.4. Advantages of MobilityAs we are working on performance based inmobile environment, the advantages of usingmobility are as shown below :4.1. Long Network LifetimeDue to the mobility we can have sensornodes moving which will result in havingtransmission more disperse and energy dissipationmore effective. Moreover when sink nodes are staticwe find that nodes near the sink get die soonercompared to other nodes. But if the nodes are mobilethen this problem can be eliminated which willdefinitely increase the network lifetime.4.2. More Channel CapacityBy creating multiple communication pathsand the number of hopes a message must travelbefore reaching the destination, channel capacity isincreased as well as data integrity can also bemaintained using mobility.4.3. Enhance Coverage and TargetingNodes are being deployed in a grid, randomor any other regions. However, an optimaldeployment is unknown until nodes start collectionof data. Deployment of nodes in remote or wideareas, rearranging node positions is generallyinfeasible. However, when nodes are mobile,redeployment is easily possible.4.4. Better Data FidelityWhen wireless channel is in poor condition,a mobile node is helpful to carry data to a destinedpoint. The reduced number of hopes will increase theprobability of successful transmission.5. Implementation DetailsWe have used a simulator Castalia 3.2,which is designed especially for wireless sensornetwork. Castalia is based on OMNET  so wewill be using omnetpp-4.2.2. The Castalia simulatorfor WSNs, for instance, provides a generic platformto perform “first order validation of an algorithmbefore moving to an implementation on a specificplatform” .The Table I will show some simulationparameters that we are going to use in order todetermine and analyze the performance of the CTPin static and mobile environment.
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804800 | P a g eTable I: Simulation ParametersParameter Value UnitNo. of Nodes 100 NodesDimension of Space 250 × 250 MetersSimulation Time 30 SecondsDeployment uniform -Mobility ManagerNameLineMobilityManager-Update Interval 100 SecondsSpeed 3meters /second6. Simulation Results and Analysis ofExisting CTPThe following are some of the importantperformance metrics that we will be using for thestudy of CTP’s performance under mobile scenarios:1. Data delivery ratio: It defines the ratio betweenthe number of data packets successfully delivered tothe sink to those sent by the source nodes.2. Control traffic: It defines the traffic resulting fromrouting beacons in the network for establishing andmaintaining the tree.3. Application level packet latency: It is the timetaken for a packet to travel from the source node tothe sink node.Now we will study some of the results that we haveobtained during our simulations. In the followingresults we will be comparing the performance ofCTP in static and mobile environment.Fig. 2. Data Delivery Ratio and Duplicate packetsAs shown in Fig. 2, the Data DeliveryRatio (DDR) with static nodes is 96.67% .It proves that CTP is very efficient routingprotocol in static environment. The ratio ofduplicate packets in static environment is 3.67%.While in mobile environment the DDR is very lesscompared to static which is just 43.33%. It meansthat when nodes are mobile very less number of datapackets are being transmitted to the destination (sinknode). The number of duplicate packets is also morein the mobile environment that is up to 6.83%. Dueto mobility the acknowledgment (of the receivedpackets) is not being received by the neighbor nodeand thus the node sends the data packet again andagain which results in increasing number ofduplicate packets.Fig. 3 and Fig. 4 shows the control packetsrequired to transmit in static and mobileenvironment. In static, the number of beaconpackets transmitted and received is very less ascompared to the mobile as shown in Fig. 3. Thenumber of transmitted packets is also more inmobile as shown in Fig.4. It shows that in mobileenvironment the control packet overhead is beingincreased.Fig. 3. Beacon packets transmitted and receivedFig. 4. Transmitted packets
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804801 | P a g eFig. 5. Application level LatencyThe end to end delay or latency is also animportant performance metric for sensor networks.As shown in the Fig.5, latency for static nodes iscompared with the nodes in mobile environment.The result shows that the average delay increasedwith the increase in node mobility.Fig. 6. Received packets breakdownFig. 6 shows that the ratio for the failure ofthe received packets is more in mobile as comparedto static environment. From the received packetbreakdown we can see that packet sending was failedbecause the signal was below the receiver’ssensitivity level. This happens when node moves outof the range of a receiver.7. SolutionWe will deploy a small number of staticnodes in our network region with all other nodesmobile. This static node will have highertransmission range compared to other mobile node inorder to have large coverage area. Due to this mostof the area in network region will be covered out.The static nodes are identified by specialflag bit set namely static(S) in the Routing Frame.Now every mobile node will have an entry of thisstatic node’s ETX in its link estimation neighbortable and routing table. The parent selection will nowbe carried out based on two criteria: static node andETX. For parent selection, a node will find out firstany static node in its neighbor and thereafter it willfind the node with the lowest ETX value. Here thenumber of retransmissions is decreased as comparedto previous CTP. The advantage of this mechanismis that it will increase the data delivery ratio withminimum control overhead. Due to some static nodesthe network will have a backup infrastructure. Thereliability of the network will also be increased as thepackets are not dropped for not finding any route dueto the changing topology with mobile environment.Proposed Algorithm:The routing algorithm for enhanced CTP isshown in the Fig. 7.The source node will find the requiredinformation from the routing table. Based on thisalgorithm, first criteria for a node will be todetermine the surrounding nodes whether there is astatic or mobile node. Thereafter, the source nodewill select its route for transmission. The source willprefer a static node first in order to transmit the data.The second criteria will be to check the static nodewith lower ETX threshold. Reliability of the data ismuch more important in the mobile scenario.Algorithm Routing Algorithm for Enhanced CTP Protocol1: Let minETX 0XFFFF2: Let minETXForStaticNode 0XFFFF3: Let Rn bestETXRoute4: Let Rs bestStaticETXRoute5: for RoutingTable[i]6: if (minETX > RoutingTable[i].ETX)7: Rn RoutingTable[i].nodeId8: minETX RoutingTable[i].ETX9: end if10: if (minETXForStaticNode >RoutingTable[i].ETX) &&(RoutingTable[i].isStatic)11: Rs RoutingTable[i].nodeId12: minETX for static node RoutingTable[i].ETX13: end if14: end for15: if (Rs != Null) && (Rs.ETX < ETXthreshold)16: SelectedRoute Rs17: else18: SelectedRoute Rn19: end ifFig. 7. Algorithm for Enhanced CTP Protocol8. Simulation Results of Enhanced CTP
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804802 | P a g eSimulation Setup:Simulation parameters are same aspresented previously except here we will useRandom Waypoint Mobility Model so that protocolis more reliable for realistic environment. In order toprove the efficiency of the proposed algorithmsimulations were run with two different number ofstatic nodes. For simulation purpose 100 nodes aredeployed randomly in a rectangular region of250×250 meters. These nodes are identical in termsof resources and energy. Based on the proposedscheme a small number of nodes are static while theremaining all nodes are kept mobile. An exampleshowing the overall scenario for network region is asshown in the Fig. 8.The nodes move in the network regionusing Random Waypoint Mobility model. The sinknode is located at (0, 0). The static nodes aredistributed using grid topology in order to cover theentire network region while the mobile nodes aredistributed uniformly. The simulation runs in twodifferent scenario: one with 9 static nodes andsecond with 12 static nodes.Fig. 8. Network Topology Example for Enhanced CTPHowever in original CTP all the nodes arekept mobile except the sink node. The minimum andmaximum speed for the mobile node is kept as 2 and5 respectively. In order cover more area transmissionrange of static node is 30m while for mobile node is10m. The transmit output power for static node is3dBm while for mobile nodes is 0dBm. Thesimulation rounds are run for different simulationtime period and each configuration, hence the resultsare aggregated.Results & Analysis:Based on the proposed scheme the resultsare emphasized more on data delivery ratio andcontrol overhead.Fig. 9. Data Delivery RatioAs shown in the Fig. 9 Data Delivery Ratio(DDR) for Enhanced CTP with 9 and 12 static nodesis compared with original CTP. The DDR ratio isbeen averaged for different simulation time period.The DDR is 54.10% for CTP original withthe same configurations. While DDR based on theproposed scheme with 9 static nodes is 75.51%.When the number of static nodes increases to 12nodes, DDR also increases to 79.59% as shown inthe Fig. 9.The number of Duplicate packets sent fororiginal CTP and enhanced CTP is as shown in theFig. 10. As shown in the figure Duplicates are higherin the Enhanced CTP as compared to the originalCTP. The reason for the increased duplicates is dueto the situation in which packet is delivered to thestatic node, but as no acknowledgement is receivedthe same packet is forwarded again. With theincreased number of static node this redundancy isalso increased. However, the reliability achieved byadding few static nodes at the cost of controloverhead is acceptable when the static nodes havehigher battery life which also increases the networklifetime.Fig. 10. Duplicate Packets
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804803 | P a g eFig. 11 shows the number of packets transmittedduring the simulation time. It is observed that thetransmitted packet ratio decreases in the EnhancedCTP as compared to original CTP.Fig. 11. Transmitted PacketsPackets transmitted for Enhanced CTP with 12static nodes are less than packets transmitted forEnhanced CTP with 9 static nodes. It shows thatcontrol overhead is decreased in the proposedscheme.As shown in the Fig. 12 the Beacon packetstransmitted and received for original CTP are moreas compared to Enhanced CTP which increase thecontrol overhead. While in Enhanced CTP with 9and 12 static nodes, the beacon packets transmittedand received are comparatively less. It shows thatcontrol overhead is reduced in the proposed scheme.Fig. 12. Beacon Packets.9. ConclusionFrom the simulation results we canconclude that the performance CTP is degraded inmobile environment. Due to the mobility topology isbeen changed frequently which results in frequenttree regeneration. Control overhead increases whenthe nodes are mobile.It can be concluded from the results that thepath metric estimation makes CTP unsuitable formobile environment. Based on this analysis anEnhanced CTP is presented to increase the efficiencyin mobile environment. Some changes wereperformed in the proposed scheme in order toincrease the performance in mobile scenario. In theproposed theory few nodes are kept static withincreased transmission range. The static nodes aredeployed in such a way that most of the networkregion is covered by them. The link estimation is stilldone but now a mobile node will select a nearbystatic node first based on the ETX value. A staticnode thereafter can transmit data to the mobile nodewith lowest ETX value.From the Simulation Results we canconclude that the Enhanced CTP is more efficientthan the original CTP in mobile environment. TheData Delivery Ratio is increased in the proposedscheme for mobile environment. The controloverhead is also reduced in the Enhanced CTP.AcknowledgmentWe would like to express our gratitudetowards our parents for their kind co-operation andencouragement which help us in completion of thework done so far.ReferencesJOURNAL PAPERS: Silvia Santini and Ugo Colesanti, "TheCollection Tree Protocol for the CastaliaWireless Sensor Networks SimulatorBasedOn Its Implementation For The," TechnicalReport Nr. 729, Department of ComputerScience ETH, Zurich, June 2011. Javad Rezazadeh, Marjan Moradi, AbdulSamad Ismail, “Mobile Wireless SensorNetworks Overview”, IJCCN InternationalJournal of Computer Communications andNetworks, Volume 2, Issue 1, February2012. S. R. Sawant, R. R. Mudholkar, V.C.PatilRajashree.V.Biradar, "Multihop Routing InSelf-Organizing Wireless Sensor," IJCSIInternational Journal of Computer Science,vol.8, no. 1, 2011 Peter PECHO, Petr HAN´ACˇEK, JanNAGY, “Simulation and Evaluation ofCTP and Secure-CTP Protocols”, RadioEngineering, Vol 19, April 2010.
Bijal Chawla, Dhaval Patel, Chandresh Parekh, Miren Karamta, Dr M B Potdar /International Journal of Engineering Research and Applications (IJERA)ISSN: 2248-9622 www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.798-804804 | P a g e K. Srinivasan, M. Kazandjieva, S.Agarwal, and P. Levis,“The beta-factor:Measuring wireless link burstiness”, InProceedings of the 6th ACM Conferenceon Embedded Networked Sensor Systems(SenSys), 2008. Athanassios Boulis, “Castalia’s UserManual”, NICTA, March 2011. J.,Filali, F.,Bonnet, C Harri, "Mobilitymodels for vehicular ad hoc networks: asurvey and taxonomy," CommunicationsSurveys & Tutorials, IEEE, vol. 11, pp. 19-41, 2009. I.F. Akyildiz , W. Su , Y.Sankarasubramaniam , E. Cayirci,“Wireless sensor networks: a Survey,Computer Networks”, The InternationalJournal of Computer andtelecommunications Networking, v.38 n.4,p.393-422, 15 March 2002. M. Chu, H. Haussecker, and F. Zhao,"Scalable Information-Driven SensorQuerying and Routing for ad hocHeterogeneous Sensor Networks", TheInternational Journal of High PerformanceComputing Applications, Vol. 16, No. 3,August 2002. W. Su, Y. Sankarasubramaniam, and E.Cayirci I. Akyildiz, "A survey on SensorNetworks," IEEE communicationsMagazine, vol. 40, 2002.WEBSITES: Athanassios Boulis et al. Castalia: ASimulator for Wireless SensorNetworks.[Online]http://castalia.npc.nicta.com.au/ OMNeT++ User Manual (Version 3.2):[Online]www.omnetpp.org/doc/omnetpp33/manual/usman.htmlTHESIS: Piergiuseppe Di Marco, “Protocol Designand Implementation for Wireless SensorNetworks”, Masters Degree Project,Stockholm, Sweden April 2008. Mahmood Ali and Sai Kumar Ravula, ”Real-Time Support and Energy Efficiencyin Wireless Sensor Networks”, MasterThesis in Computer System Engineering,School of Information Science, Computerand Electrical Engineering, HalmstadUniversity, Technical report, IDE0805,January 2008 Ays¸eg¨ul T¨uys¨uz Erman, “Multi-SinkMobile Wireless Sensor Networks:Dissemination Protocols, Design andEvaluation”, Degree of doctor, Universityof Twente, September 2011CONFERENCES: S. Singhal, A.K. Gankotiya, S. Agarwal,and T. Verma, ”An investigation ofwireless sensor network: A distributedapproach in smart environment”, InAdvanced Computing CommunicationTechnologies (ACCT), 2012 SecondInternational Conference on, pages 522 –529, Jan. 2012.BOOKS: E Holger Karl and Andreas Willig.Protocols and Architectures for WirelessSensor Networks. John Wiley & Sons, Ltd,Chichester, UK, April 2005. ISBN9780470095126.doi: 10.1002/0470095121.[Online] URL:http://doi.wiley.com/10.1002/0470095121.