I. the proposed algorithm. The deployed code
I. RESULTS AND EVALUATIONTo implement this algorithm, I have used Cooja -2.7-simulator, which is a sensor network simulator runs on Contiki OS.Implementation requires three major components to be set: (a) Server nodeswhich are responsible for sending update messages to the RD, (b) RPL borderrouter; a bridge between server nodes and the resource directory, and (c)Resource Directory which is a server keeps listening on a certain port forupdate messages.
However, to implement (a) and (b) I have used ‘Sky mote’sensors. The deployed code on (a) is an implementation of a simple resource,energy consumption using Joule’s law and ‘Energest’ function -which will bediscussed later in this paper- for each module, and finally, ‘update messages’code that initialize and send those messages to the Resource Directory everynumber of seconds according to the proposed algorithm. The deployed code on (b)is the default border router code ‘rpl-border-router.
c’. We need to connect theCooja border router to the external network outside Cooja See Fig. 2. Fig. 2. Connecting Cooja border router to theexternal work To implement (c) I have used an external JAVAframework called Californium (cf); a JAVA Implementation of CoAP clients,servers and resource directory.
I only used the resource directory executablejar. Now, the communication between CoAP server nodes in Cooja and the resourcedirectory needs the border router to have a place in our network. However, tostart the Resource Directory, we only need to run the executable jar called(cf-rd-1.1.0-SNAPSHOT), See Fig. 3. Fig.
3. Starting the Resource Directory server usingCalifornium JAVA implementation After all, these components are ‘up and running’ andaccording to the fact that the new enhanced algorithm is based on the batterylevel of each sensor, we need to find out a way to calculate the residualenergy in a certain node. Unfortunately, there is no direct way to get thatvalue when using Cooja. Instead, Contiki OS has a built-in function calledEnergest. Using this function, you can measure the power consumption fordifferent modules by having the time the components have been turned on.
Components are CPU; CPU in active mode, LPM; CPU in idle/low power mode, RX;Radio in receiving mode and finally TX; Radio in transmitting mode.Substituting the time each component has been turned on in Joule’s law (seeEqu. 1,2,3 and 4) will yield the energy consumed by that component. Summationof components energy consumption will result in the total power consumption bya certain node. Finally, if we substitute the total power consumption of acertain node in Equ.
5 5, we will have the Battery Level at a certain pointin time 5. (1) (2) (3) (4) (5)TPC: Total Power Consumption,evaluated by summation of equations (1-4)FullBattery: 7000 mj(assumed)I_x: Current value in xmodule from the data sheet.V_x: Voltage value for modulex from the datasheetRTIMER_SECOND: Number ofTicks Per Second in Contiki Table 1 shows thesimulator environment variables (same values as in 5) To evaluate the new algorithm fairly, we need to compare it to theStandard algorithm and to the traditional Dynamic one as well. After doing manyexperiments that discovers many aspects such as a different number of nodes,different RX ratios and different periods of simulation time, some interestingresults have been discovered.
In general, the new algorithm has extended the networklifetime by 10% over the traditional dynamic approach 5 and 35% over theStandard one. Fig. 4 shows thebattery level over the lifetime of the three approaches (Standard, Dynamic andenhanced). Obviously, according to the lifetime, Enhanced dynamic approachoutperforms the other two approaches for all battery levels.Moreover at 5%, andwhen the battery is nearly dead, the Enhanced approach has made it 10% morelife than the traditional Dynamic one. Fig.4. Battery Level over the time for Standard, Dynamic and Enhanced Taking another measurement into consideration, I have trieda different number of nodes to evaluate how network density would affect thenew algorithm.
Fig. 5 shows that the new Enhanced dynamic algorithm has given abetter life for a various number of sending nodes. For example, when trying 70nodes, the new technique has given almost 100 min more lifetime over theStandard approach and 30 mins over the traditional Dynamic technique. Fig. 5.
Lifetime for variousnumber of nodes The last Measurement I took into consideration is trying a differentreceiving ratio (RX), which is a random variable that allows packet receivingwith random errors 15. I have run the simulation om different RX ratios (75%,50%, 25% and 0%), and as Fig.6 shows again the excellence of the enhancedapproach over the previous approaches. Fig. 6.
Network lifetime over Different RX ratiosII. CONCLUSION AND FUTURE WORK In this project, I proposed an enhanced dynamic algorithm that extendeda previous effort of 5, I have increased the update interval of each batterylevel as mentioned before to reduce the overall number of sent messages thathave been costing more energy, this applies to less-sensitive application thatrequires a very frequent update messages to guarantee the refreshment of theirsensitive data.However, this new algorithm has proven an extended network lifetime by10% over the Enhanced dynamic one.On Future, I plan to test this algorithm with a denser network toverify the maximum improvement this technique can provide.
I plan to study as well, this new technique in a larger area than (200x 200) to know the impact of the sensor field space. References1Somayya Madakam, R. Ramaswamy and Siddharth Tripathi”Internet of Things (IoT): A Literatur Review”. 2 K.Matthias, Martin Lanter, and Zach Shelby.
“Californium: Scalable cloud services for the internet of things withcoap.” Internet of Things (IoT), 2014 International Conference on the.IEEE, 2014.3 L.Meirong “Distributed resource directory architecture in Machine-to-Machinecommunications.” WiMob. 2013.
4Pötsch, Thomas, et al. “Performance evaluation of CoAP using RPL and LPLin TinyOS.” 2012 5th International Conference on New Technologies,Mobility and Security (NTMS). IEEE, 2012.5Muneer Bani Yasin, Qusai Abuein, Ahmad Bani Amer, and Mamoun Qasem.
“AnEnergy-efficient Technique for Constrained Application Protocol of Internet ofThings.”6K. Matthias, Simon Duquennoy, and Adam Dunkels. “A low-power CoAP forContiki.
” 2011 IEEE Eighth International Conference on Mobile Ad-Hoc andSensor Systems. IEEE, 2011.7C. Walter, Kris Steenhaut, and Niccolò De Caro. “Integrating wirelesssensor networks with the web.” Extending the Internet to Low power andLossy Networks (IP+ SN 2011) (2011).8 Shahid Raza,HosseinShafagh, and Kasun HewageLithe:Lightweight Secure CoAP for the Internet of Things