Proceedings of the International Conference on Big Data, Cloud and Applications Tetuan, Morocco, May 25 - 26, 2015 Optimization energy consumption in mobile cloud computing by using an elastic framework Yassine ELMAHOTI Noura Aknin Information Technology and Modeling Systems Research Information Technology and Modeling Systems Research Unit, LIROSA Laboratory Unit, LIROSA Laboratory Faculty of Science, Abdelmalek Essaadi University Faculty of Science, Abdelmalek Essaadi University Tetuan, Morocco Tetuan, Morocco aknin@ieee.org yassinemahoti@yahoo.fr Abstract— The mobile cloud computing (MCC) has become constant mobility of mobile which contribute significantly to a more and more present in our life and that is due to the wide quickly landfill of batteries availability of mobile devices in the world market (smartphones, tablets, etc.), however unlike the cloud computing has proven a This paper propose a an implementation of a framework for big success and performance in communication technologies, the dividing a customer request between a number of devices (best mobile devices are not able to a fully benefit from this elected), this electing is based on the following criteria:( development due to their resource’s limitations such as the CPU, distance to the mobile concerned, the load of the battery, the Ram, Battery life, etc. utilization of CPU, RAM, storage, power signal), the obtained Today, due to heavy applications that are hosted in the cloud, results showed that the installation of this framework the mobile devices consumes more and more energy, and it contributes greatly to save energy to consumption. contributes greatly to the battery discharge in a very short The paper is organized as follows: Section (1) gives an delays. overview to the mobile cloud computing, Section (2) describes To resolve this problem, we propose in this paper an the framework architecture, its algorithms and its results. implementation of an elastic framework for splitting a customer request between the others devices (best elected devices) for II. MOBILE CLOUD COMPUTING minimizing the processing time, the results have shown that the Facebook, Whatsapp, YouTube, are the most application used implementation of this framework contributes significantly to a by million people in the word, for example Facebook users considerable minimization of the request processing time and have arrived at 1,393 milliard connected in January 2015. therefore minimizing the energy consumed. This explosion in number of users is surely due to the large Keywords— mobile cloud computing, energy, framework, distribution of mobile devices in the word, a study conducted agent by Gartner (June 2014) showed that the number of Worldwide Device is reach 256 Million Units as shows the table I: I. INTRODUCTION TABLE I. WORD DEVICE UNIT The use of mobile devices has seen a great explosion of the last Worldwide Device Shipments by ten years and have become an indispensable way in our daily (Segment Thousands of Units) 2013 2014 2015 life for checking information in the internet; a user can now Traditional PCs (Desk-Based and view his inbox anywhere and anytime while before it needs to Notebook) 296,131 276,221 261,657 be sitting at his desk for getting this information. Ultra mobiles, Premium 21,517 32,251 55,032 The mobile cloud computing has proven a big success by Tablets 206,807 256,308 320,964 permitting a huge facility and flexibility to give an information Mobile Phones 1,806,964 1,862,766 1,946,456 faster and a powerful additional resources for devices that can run a very heavy applications (CPU, RAM, storage as a service). Despite that, the mobile cloud computing (MCC) knows some problems and especially the high energy consumption problem and this is due to several factors: complexity of application, the strength of the wireless signal, distance to the base station, the 51 heavy processing (image processing, software localization, Worldwide Device Shipments by Segment etc.), it permit also a remote storage resources and virtual 2 000 Traditonal PCs networks to connect to remote application. 1 800 (Desk-Based and 1 600 Notebook) PaaS (platform as a service): it is another model for the 1 400 delivery of cloud computing services, allowing application Ultramobiles, 1 200 developers to prepare libraries and prerequisites needed to Premium program, test their applications in secured and reliable 1 000 environment. 800 Tablets 600 SaaS (software as a service): the user does not need to install 400 any tool, the software and data are stored in cloud providers, 200 Mobile Phones and the user through a web browser can connect to the service to do the desired operations. 0 2013 2014 2015 The connection of users to the cloud computing service is provided by wireless networks (3G, 4G, WiFi) as shown in the figure below, and this type of connection sometimes generates Fig. 1: Worldwide device units a fast discharge battery because of the weakness the signal, far distance between the base station and the mobile receiver, The Fig. 1 shows this recent years an important number of users have emerged from the traditional computer to mobile devices because their performance become more efficient and sufficient for doing their majority tasks in internet, for example the SAMSUNG galaxy S5 have 2gb in memory, 2,5 GHz quad core of CPU and it capable to execute a several process in the same time. This significant emerging from the fixed equipment to the mobile is due to the possibility of mobile equipment has exceeded the old and simplest communication tools (call / message) and can now guaranteed other interesting things service browsing in internet, install applications, share data, etc. The cloud computing recently appeared in the IT word allows users to execute applications or store data without having the necessary resources in their terminals, the entire treatment of task is guaranteed by the cloud computing servers which are robust and powerful. Fig. 3: mobile cloud computing architecture The multi task operation ensured by smartphones was The remarkable development of cloud computing in recent introduce the cloud computing in the mobile environment, a years, attracting more and more interest from various internet simple user with his terminal can benefit from a several IT users and IT looking to enjoy the best services and applications services as a model as shows the figure below: available online through the web. This is a new business model that cloud computing promises to ICT. Indeed, the model promises a change in the mode of investment and operation of IT resources. III. PROPOSED WORK To resolve to the energy consumption problem for the mobile devices, we propose an implementation of a framework which is composed by several subsystems interconnected between them for a smart processing of the customer queries. This framework permits a splitting request of the cloud client between the idlest mobiles available in the network for Fig. 2: Cloud computing models minimizing the processing time and therefore saving the energy consumption for executing a concerned process. - IaaS(infrastructure as service): it is the lower level of cloud computing services, it permit to the client to benefit of hardware resources like as a remote RAM and CPU to execute 52 a) Description This framework is as an intermediate software layer between and came back the result to the customer in a very short time the mobile terminal and cloud providers: it receives the than the normal case by splitting the request by many others request from the client, communicates with the cloud servers mobiles devices . This framework is composed by Fig. 4 : The Elastic framework - Mobile agent: is an agent installed in the mobile for receive c) Algorithm and send information to the other component of this framework. The processing time of request client depends to the transfer data flow from the cloud providers to the terminal mobile and - Job stat agent: It returns for the Job splitter agent the stat of its resources CPU, RAM, so we have: the request processing and inform it when the operation if fails for choosing another device. proc _ time  computation _ time  transfer _ time (1) The main elements of this framework: With proc _ time is the processing time for executing a - Localizer agent: localize the mobile devices which are near tasks, computation _ time is the total time required for to the client concerned. compute the tasks by the mobile, transfer _ time is the total - Resource monitor agent: It returns the resources allocation of time for the transfers the data from the providers cloud to the mobiles devices that have be mentioned by the localizer agent mobile. (CPU, RAM, buttery life, storage, signal power). Data - Job splitter agent: it splits the client request between the best computation _ time  (2) elected mobile (less overloaded) returned by the resource CPU monitor and store in his cache memory the state of the process With CPU is the processor of the mobile executed in the mobiles devices. Data transfer _ time  (3) Bandwidth1 53 With bandwidth is data flow from the cloud providers to the calculating the power usage of the mobile devices, we mobiles of between the mobiles theme self. installed a software called Joulemeter a tool developed by Microsoft researchers and we found the following result as From (2) & (3) we have show the table I and the figure: Data Data proc _ time   (4) CPU Bandwidth1 Power Usage by number 1 1 25 proc _ time  Data(  ) (5) of devices CPU Bandwidth1 20 When we introduce the elastic framework the task client is 15 1 divided by the best elected others mobiles. 2 1 1 1 10 3 proc _ time  Data ( n  ) 1  CPU n Bandwidth1 4 5 n i 1 (6) Data 0  Power Usage (Watt) n.Bandwidth 2 Where Bandwidth1is the bandwidth between the BTS and Fig. 6: Power usage by number of device the mobile, Bandwidth2 between the mobile elected and the mobile concerned and n is the number of selected devices by the framework. To ensure that the energy consumed by a mobile for d) Result processing the request’s customer in the case without using the framework exceed the energy summation by all selected To experiment our algorithm proposed in this paper, we set up devices by the framework we calculated the energy by the a cloud platform composed of 3 servers and 10 mobiles following formula: devices with each one have 1.6 Ghz in its CPU and we found this following this results: n Energy  proc _ time *  power _ usagei i 1 3000 Processing time by number of In case of introducing the elastic framework the power usage 2500 devices come the sum of power of the different devices elected: n power _ usage   power _ usagei 2000 1 1500 2 i 1 3 Where n is the number of selected devices by the framework 1000 4 So the energy consumption becomes: 500 n 0 Energy  proc _ time *  power _ usagei Processing time (s) i 1 Fig. 5: The processing time We have the following result : As we can see, when we introduce the elastic framework, the processing time decreases in a very meaningful way and this is because the response of the request client from the cloud is 80000 divided by the selected mobile those respond to the following 60000 1 criteria (CPU utilization is under 50% RAM usage is under 50%, the distance between this mobile and the client 40000 2 concerned is closer than the client with the BTS, the signal 3 strength and the rate of the battery that exceeds 50%). 20000 4 As shown the figure 6, this framework not only decrease the 0 processing time, but also decreases the power usage and this is Energy cosumption(Joule) due to the CPU came more and more offload and therefore the mobiles use a little power for processing the task. For 54 Fig. 7: Energy consumption by number of device [6] Ricky K.K. Ma, Cho-Li Wang, “Lightweight Application-level Task Migration for Mobile Cloud Computing .” In Proceedings of 26th IEEE After this result obtained, we can say the implementation of International Conference on Advanced Information Networking and this framework permits for client of mobile computing to Applications, 2012 minimize the response time with a very low energy [7] L. Yang, J. Cao, S. Tang, Tao Li, Alvin T. S. Chan, “A framework for Partitioning and Execution of Data Stream Application in Mobile Cloud consumption. Computing.” IEEE Fifth International Conference on Cloud Computing, 2012. IV. CONCLUSION [8] Rellermeyer, J.S., Alonso, G., Roscoe, T.: R-OSGi: Distributed applications through software modularization. In: Proceedings of the The autonomy of batteries has become a serious problems for ACM/IFIP/USENIX 8th International Conference on Middleware any person use a device mobiles, the lifetime of batteries has (Middleware’07). Volume 4834 of LNCS., Springer (2007) 1–20. become very short due to the heavy applications hosted in the [9] M. T. Nkosi and F. Mekuria, “Cloud Computing for Enhanced Mobile mobile cloud computing providers. Health Applications,” in Proceedings of the 2nd IEEE International Conference on Cloud Computing Technology and Science, pp. 629, The utilization of the elastic framework proposed in this paper February 2011. allows users to keep theirs batteries life longer by minimizing [10] M. T. Nkosi and F. Mekuria, “Cloud Computing for Enhanced Mobile the processing time of their demands and offloading their Health Applications,” in Proceedings of the 2nd IEEE International resource’s equipment allocations. Conference on Cloud Computing Technology and Science, pp. 629, February 2011. Thanks to this framework, the mobile cloud computing will [11] Jayant Baliga, Robert W. A. Ayre, Kerry Hinton, and Rodney S. Tucker, become more robust, efficient and friend of environment “Green Cloud Computing:Balancing Energy in Processing, Storage, and because it offers a treatment in a very short time for a client Transport”. Vol. 99, No. 1, January 2011 | Proceedings of the IEEE task with a very minimal energy consumption and therefore the client wins some additional energy that he can use for [12] Koshy, K.I.; Juby, A.M.; Namboodiri, V.; Overcash, M.; , "Can cloud other applications. computing lead to increased sustainability of mobile device?," Sustainable Systems and Technology (ISSST), 2012 IEEE International However, it should be noted that the volume of the experience Symposium on , vol., no., pp.1-4, 16-18 May 2012 remains insignificant to draw a generalized conclusions. It is [13] Scott Paquette, Paul T. Jaeger, Susan C. Wilson, A research paper on desirable for validate our proposals work expand the tests on a Identifying the security risks associated with governmental use of cloud computing, Govt. Information Quarterly, Volume 27, Issue 3, July 2010, large scale of use (thousands of virtual machines and mobile Publication Elsevier, Pages 245-253. users) [14] Byung-Gon Chun, Sunghwan Ihm, Petros Maniatis, Mayur Naik, and Ashwin Patti. Clonecloud: Elastic execution between mobile device and REFERENCES cloud. In Proc. of EuroSys, 2011 [15] M. S. Gordon, D. A. Jamshidi, S. Mahlke, Z. M. Mao, and X. Chen. COMET: code offload by migrating execution transparently. In USENIX OSDI, pages 93–106, 2012. [1] X. Gu, K. Nahrstedt, A. Messer, I. Greenberg, and D. Milojicic, “Adaptive offloading inference for delivering applications in pervasive [16] Y. Kwon, S. Lee, H. Yi, D. Kwon, S. Yang, B.-G. Chun, L. Huang, P. computing environments,” in Proc. of the 1st IEEE International Maniatis, M. Naik, and Y. Paek. Mantis: Automatic performance Conference on Pervasive Computing and Communications prediction for smartphone applications. In USENIX ATC, 2013.. (PerCom’03), Fort Worth, Texas, USA. IEEE, March 2003, pp. 107– [17] P. Shankar, B. Nath, L. Iftode, W. Huang, and P. Castro. Crowds replace 114. experts: Building better location-based services using mobile social [2] L. Liu, R. Moulic, and D. Shea, “Cloud Service Portal for Mobile network interactions. In Proc. of IEEE PerCom, 2012. Device Management,” in Proceedings of IEEE 7th International [18] H. T. Dinh, C. Lee, D. Niyato, and P. Wang. A survey of mobile cloud Conference on e-Business Engineering (ICEBE), pp. 474, January 2011. computing: architecture, applications, and approaches. Wireless [3] S. Kosta, A. Aucinas, P. Hui, R. Mortier, and X. Zhang, “ThinkAir: Communications and Mobile Computing, 2011. Dynamic resource allocation and parallel execution in the cloud for [19] Peoples C, Parr G, McClean S, Scotney B, Morrow P (2013) mobile code offloading,” in Proc. of the 31st IEEE International Performance evaluation of green data centre management supporting Conference on Computer Communications (INFOCOM’12), Orlando, sustainable growth of the internet of things. Simul Model Pract Theory Florida, USA. IEEE, March 2012, pp. 945–953. 34:221–242. [4] N. Fernando, S. W. Loke, W. Rahayu, “Mobile cloud computing: A [20] S. Xavier and S. J. Lovesum, "A survey of various workflow scheduling survey”, Future Generation Computer Systems, vol. 29, no. 1, Jan. 2013, algorithms in cloud environment," International Journal of Scientific and pp. 84–106 Research Publications, 3(2), 2013. [5] M. T. Nkosi and F. Mekuria, “Cloud Computing for Enhanced Mobile Health Applications,” in Proceedings of the 2nd IEEE International Conference on Cloud Computing Technology and Science, pp. 629, February 2011. 55