Efficient Partitioning and Offloading Tasks in Edge Computing by Fruit Fly Optimization algorithm

(1)

ISSN: 2005-4238 IJAST

265

Efficient Partitioning and Offloading Tasks in Edge Computing by Fruit Fly Optimization algorithm

Kulvir Singh

¹

; Dr. Yogesh Kumar Sharma

²

1Research Scholar Department of Computer science and Engineering; Shri JJT University Jhunjhunu Rajasthan

2Associate Professor (HOD/Research coordinator), Department of Computer science and Engineering; Shri JJT University Jhunjhunu Rajasthan

Abstract—We explore a nature of experience (QoE) based calculation offloading booking issue for Edge Computing, in which information preparing and basic leadership are put at the edge of the Internet and near keen cell phones and end clients. Taking into account that shrewd gadget proprietors esteem both reaction time and battery life, it is sensible to appropriately address the inertness and vitality tradeoff. This paper catches a client driven view to handle the offloading planning issue by means of mutually assigning correspondence and calculation assets with thought of the QoE of clients. We define our structure as half breed Fruit Fly enhancement calculation and unravel it in a productive manner. Numerical results exhibit that the proposed offloading plan accomplishes an improved presentation on dormancy time and vitality utilization, when contrasted with benchmark plans.

Keywords— Artificial Intelligence, Computation Offloading, Edge Computing, Task Migration I. Introduction

As of late, the promotion of cell portable correspondence and quick improvement of 5G innovation make brilliant cell phones, (for example, cell phones, wearable gadgets [1], keen autos, and so on.) generally show up in individuals' everyday life. The expanding portable traffic and muddled registering bring extraordinary difficulties to organize correspondence and processing assets. In the previous barely any decades, extraordinary advancement has been made in remote correspondence and distributed computing innovations. Be that as it may, because of elements, for example, inadequate figuring and capacity limit of cell phones and the constrained battery limit brought about by equipment structure, neighborhood processing can just cover a couple of basic calculation assignments. Figuring serious and information escalated assignments can just depend on uplink offloading to the cloud for handling [2]. This implies the insufficiency of nearby gadgets in these perspectives are helped by distributed storage, registering and correspondence assets. Such an example works when the quantity of clients is little or the sort of utilization is straightforward.

Be that as it may, in a multi-client mode where huge information are created, the transmission procedure of calculation errands from cell phones to the cloud possesses a great deal of system correspondence assets.

Fig. 1 Offloading in Edge Computing

It will effectively surpass the farthest point of system security load, lead to arrange clog and cause unsatisfactory correspondence delay [2]. Accordingly, conventional distributed computing offloading mode isn't appropriate for postpone delicate calculation errands in 5G period. So as to adapt to these calculation assignments with low deferral, high multifaceted nature and high dependability prerequisites [3], we have to depend on further developed processing modes.

Edge registering is another processing mode proposed by ETSI MEC ISG (Industry Specification Group) made out of 6 individuals including Nokia and Huawei [4]. Presently it has become the main hypothesis and calculated structure supporting 5G inquire about.

Through sending of server on the edge of web (counting the remote passageway (AP), switches, switches, base stations, server farms or cloud stages and some other gadgets with capacity and registering limit), distributed computing and distributed storage are pushed by edge processing in places nearer to portable clients, in this manner to give 5G administrations to cell phones.

(2)

266

Fig. 2. Structural diagram of MEC.

Be that as it may, regardless of whether edge figuring is received as the corresponding method of distributed computing, in the period when 5G versatile correspondence goes into commercialization, the simultaneous handling of substance solicitation and calculation task still causes impressive weight on savvy correspondence types of gear [5]. (1) The calculation assignments from shrewd gadgets are differentiated. At present, there are numerous kinds of shrewd applications. In any case, these applications include diverse calculation assignments and information types, including organized information, for example, advanced sign and gigantic unstructured information, for example, content, pictures, sound and video. What's more, these entangled calculation assignments can frequently be isolated into various subtasks for parallel preparing. (2) Network assets are powerfully evolving.

Processing administrations in MEC engineering are obliged by countless questionable factors, for example, the quantity of portable clients (calculation errands), arrange framework security, correspondence and figuring asset booking strategy. The most effective method to constantly furnish administrations with high unwavering quality and low postponement for clients through joint improvement of above factors is a trouble issue in ebb and flow MEC inquire about. (3) The figuring limit of edge cloud is limited and heterogeneous. At the point when the assignment is offloaded to edge cloud, it might experience the issue that the computational intricacy surpasses the registering limit of edge cloud. Right now, the calculation task offloading methodology that adjusts to dynamic change of processing assets should be considered. Additionally, contrasted and the conventional cloud server, the edge server's processing limit isn't sufficient to help a wide range of calculation errands. In this manner, it is important to extensively consider the issue of calculation relocation when the traffic information of undertaking change progressively.

Fig. 3. Selecting mobile edge server to offload computation.

In this way, it is a pressing issue to utilize constrained and heterogeneous figuring assets tense cloud, to plan a streamlined insightful calculation offloading system and to diminish the preparing delay. So as to improve the exhibition of offloading methodology, we have to assess the information volume of assignments ahead of time, in order to figure the ideal offloading technique at the first run through when a versatile client demands administrations. Artificial intelligence is presently a propelled hypothesis and

(3)

267

innovation in the field of information preparing and investigation. As a part of software engineering, profound learning innovation empowers us to direct inside and out handling and mining of an assortment of unstructured information gathered from cell phones (particularly when there is a lot of recorded information), subsequently to give increasingly astute psychological administrations to MEC framework [6]. As the latest ideal forecast innovation, edge processing starts to utilize profound learning calculations, for example, RNN [7], GRU and FRUIT FLY OPTIMIZATION ALGORITHM [8] to give intellectual capacities to organize administrations, traffic, load and other framework exhibitions to upgrade nature of administration (QoS) and nature of experience (QoE). Likewise, edge hub joined with AI innovation is required to advance the improvement of edge processing by giving conveyed AI administrations.

The principle commitments of this paper are as per the following:

• A new canny calculation offloading based MEC engineering is structured from two parts of framework and business rationale.

• By coordinating AI, neighborhood figuring and edge registering, a calculation task expectation calculation dependent on FRUIT FLY OPTIMIZATION ALGORITHM is proposed for MEC design.

• An ideal calculation offloading procedure dependent on task forecast is proposed for cell phones.

• In perspective on the limit and heterogeneity of processing tense cloud, a calculation task relocation for edge cloud booking plan is advanced to help with enhancing the offloading system.

II. Related work

An enormous number of existing works have examined the advancement issues on versatile edge processing. The greater part of these works center around the basic leadership issue for task offloading at the client end. These works will initially assess the normal deferral or vitality utilization for target errands and afterward offload them to the versatile edge if the postponement or vitality utilization is decreased. As indicated by the improvement objective, we partition these works into two classifications: chips away at upgrading vitality utilization and takes a shot at diminishing the undertaking delay. Our work falls into the subsequent classification and contrasts from the current works in that our point is to decrease the deferral at the system end rather than the client end.

Takes a shot at improving vitality utilization. K. Zhang et al. [17] exhibited a multi-gadget calculation offloading system for MEC and figured an enhancement issue that limited the gadget vitality utilization. A three-organize offloading plan was proposed to acquire the imperfect arrangement, which (1) ordered the cell phone, (2) decided the need and (3) apportioned the radio asset. X.

Chen et al. [9] further considered the obstruction and impacts when there were such a large number of clients attempting to offload errands to similar sBS, which can fundamentally build the vitality utilization of cell phones. The offloading was planned as a multi- client game, which was demonstrated continually conceding a Nash harmony. W. Labidi et al. [18] considered the time differing channel state for remote offloading and proposed a planning plan for task offloading, which attempted to utilize remote channels and client cushions to diminish the vitality utilization.

Deals with diminishing the assignment delay. J. Liu et al. [11] attempted to limit the execution delay for single clients with one- dimensional hunt calculation. The calculation yielded an arrangement for offloading choice as indicated by the application support lining state. Additionally, the qualities of remote channels were likewise considered. Y. Sharma et al. [35] together advanced the undertaking offloading booking also, transmitted power allotment issue to diminish the offloading delay. P. Chahar et al. [11]

considered the spatial assorted variety of the sBSs in the offloading procedure. The sBS that was liable for task execution was picked by the clients. At that point the outcomes would be come back to the clients through the sBS with the most noteworthy RSSI of the remote connections. Be that as it may, the work was intended for single assignment offloading.

There are additionally a few works that mutually streamline both vitality and assignment delay. Lessening task delay regularly includes extra vitality utilization in MEC, particularly for the assignments that execute quicker on cell phones than on versatile edges. A few works set a vitality limit and afterward limit the execution delay without surpassing the vitality edge. For instance, Y. Mao et al. [14] proposed a dynamic offloading plan for single client to limit the execution delay for vitality gathering gadgets, where the vitality reaping procedure added unpredictability to the offloading calculations.

A lightweight guess approach was proposed to accomplish a decent tradeoff among multifaceted nature and the defer minimization.

U. Rani et al. [20] utilized a multi-arrange successive game model to meet the vitality and postpone necessities simultaneously.

Not quite the same as the previously mentioned works, our work accentuates on limiting the postponement inside the MENs. The postponement in MENs not just adds to the general undertaking execution delay yet additionally impacts the basic leadership process at the client end. We propose a novel in-organize task planning way to deal with diminish the undertaking execution delay, where task data, asset data and client versatility are mutually considered. Not quite the same as the [13] which utilized the contact rate to catch client portability, our work is based over the client direction forecast [15,16], which can precisely catch client versatility in most MEC situations, for example, air terminals, shopping centers, and so forth. Also, to manage the effect on offloading choice, we propose a novel postpone estimation plot for cell phones.

III. System architecture

So as to manage issues of enormous information cooperation, correspondence assets deficiency and extreme postponement of calculation offloading in distributed computing mode, we propose another MEC design dependent on clever calculation offloading, as appeared in Fig. 2. This engineering is separated into two sections: framework (left part in Fig. 2) and business rationale (right part in Fig. 2).

The foundation can be separated into three sections: (1) Mobile gadgets every day, including advanced mobile phone, tablet PC, selfdriving autos, wearable gadgets and robots [16]. These gadgets can give an assortment of utilizations and administrations, depending on progressively ground-breaking equipment offices, yet additionally on the AI calculations gave by the foundation framework. In the conventional distributed computing mode, nearby gadgets discuss straightforwardly with the cloud server, gather

(4)

268

client information locally and send calculation assignments legitimately to remote cloud. Notwithstanding, this mode would cause channel blockage of access organize under enormous information cooperation, which isn't helpful for give delay-touchy canny administrations. Subsequently, the edge haze of the center layer is inferred between neighborhood gadgets and cloud in MEC design, which is utilized to aid correspondence and registering. (2) This part incorporates servers conveyed anxious of web, (for example, remote passageway, base station and switches), and these servers are likewise called as edge registering hubs. These hubs can speak with nearby cell phones through remote channel, share a few errands offloaded by clients with their restricted figuring assets, and move muddled calculation assignments to remote cloud for further handling through fronthaul connect. (3) With amazing registering and capacity assets, cloud servers can give AI administrations. The blend of distributed computing and AI is alluded as the foundation of psychological registering. It can compensate for the deficiency of edge distributed computing in knowledge, attempt increasingly muddled calculation assignments, send the outcomes back to edge hubs by means of backhaul, lastly send them to versatile clients through remote system.

Business rationale is likewise partitioned into three sections, relating to those of foundation: (1) Due to the restricted processing and capacity limit of nearby gadgets, versatile clients can just deal with some straightforward calculation undertakings locally.

Progressively entangled errands would be offloaded to edge cloud through remote channel. (2) In thorough thought of the anticipated assignment size, hub registering assets, hub vitality hold and different factors, edge figuring hubs will choose whether this undertaking ought to be prepared locally or moved to different hubs. (3) Service substance are frequently legitimately sent through remote cloud for increasingly entangled applications. Some figuring hubs likewise plan calculation assignments to the cloud for preparing. This ensures administration knowledge to the detriment of correspondence delay. The terminal layer, get to layer and system layer in framework relates to these three sections individually.

The reason and favorable circumstances of planning such another MEC design dependent on astute calculation offloading are as per the following: (1) It is appropriate for heterogeneous information and expanded calculation assignments. Expectation for various undertakings is a key advance in calculation offloading. In the event that the sort, size and registering assets necessities of calculation errands can be anticipated ahead of time, the basic reference premise can be accommodated the enhancement of offloading technique. (2) It is versatile to progressively changing system correspondence assets and calculation errands. Because of dynamic changes in organize assets and calculation undertakings, the enhanced offloading technique can help to improve QoE from different points of view, for example, computational postponement and unpredictability. The streamlined errand movement can likewise diminish clog in get to arrange. (3) It can make up the deficiencies of nearby gadgets and edge processing hubs in registering and capacity limit. The MEC engineering dependent on wise errand forecast and calculation offloading can assist nearby gadgets with handling progressively confounded calculation and assuage the weight on remote cloud.

IV. Task dependency model

The configuration of the MEC for the most part incorporates three sections: the cloud server farm layer, the edge server layer, and the cell phone layer. the cell phone layer incorporates sensors, workstations, cell phones and different gadgets with low handling execution; At the point when the cell phone needs to improve execution through assignment offloading, an entire versatile application will be first separated into a few subtasks by some division calculation. They are normally information handling assignments with a lot of calculation. The isolated subtasks have information association with one another.

Every hub si ∈ S of the diagram speaks to the partitioned subtask, and each edge buv ∈ B of the chart speaks to the change information between undertakings. For instance, bij shows that the assignment sj must get the preparing aftereffect of undertaking si before it can keep on executing. the arrangement of subtasks in the wake of isolating a versatile application is S ={s1, s2, s3, s4, s5}, where s1 and s5 must be executed on the neighborhood gadget, and remaining subtasks can be offloaded as required.

Simultaneously, coordinated bolts are utilized to show information reliance among subtasks. For instance, the s5 must get the information of s2 and s4 before it can keep on executing.

V. Fruit Fly Optimization Algorithm in Efficient Computation Offloading

As conventional distributed computing offloading and edge registering offloading modes just consider the direct offloading of calculation errands, their offloading system can't guarantee the base deferral. So it isn't insightful enough. Thusly, the offloading system of edge registering will be advanced and improved in this paper from three viewpoints: (1) Computation task forecast calculation dependent on Fruit Fly Optimization Algorithm. The profound learning calculation is acquainted with foresee the highlights of assignments and to help the judgment of figuring delay in offloading system. (2) Computation offloading system for cell phones dependent on task forecast.

After Fruit Fly Optimization Algorithm calculation is utilized to get increasingly precise information of undertaking traffic, far reaching assessment will be led to offloading execution as per different highlights of edge distributed computing hubs to get ideal offloading methodology. (3) Computation task movement for edge cloud booking plan. So as to additionally lessen processing delay based on streamlined calculation offloading procedure, another errand relocation conspire is acquainted with help with planning between edge mists. The stream outline for in general calculation is as appeared in Fig. 3.

(5)

269

Fig 4 Flow chart of Fruit Fly Optimization Algorithm VI. Performance evaluation

To look at the productivity of the proposed structure, we plan numerous assessments so as to answer the accompanying questions:

(1) is the normal cost model approved enough? (2) does the structure strong to various framework states? what's more, (3) the productivity of flaw tolerant system? Every datum point is rehashed multiple times for every application recorded in Table 3 to get the normal esteem.

Effect of cost estimation deviation

So as to demonstrate that the internet learning structure is approved in inferring an ideal procedure. In this analysis, we look at how the standardized normal cost changes as the assignment number develops on all the three application assignments recorded in Table 3. The online calculations are contrasted with disconnected calculations with show the approval of evaluated cost model and what we concern is the means by which the presentation of online calculations improve in the procedure of errand offloading (the control approach got from on the web calculation can refresh and adjust to the ever-changing portable setting in principle). Accordingly, disconnected calculations are utilized as ideal ones and the assessment thinks about how the presentation of on the web calculation to draw near to the ideal one. For every application, the assessment is rehashed multiple times lastly get a normal cost for all the three calculations. The normal expense is analyzed for the case with deficiency tolerant system and without issue tolerant system.

Fig. 5. Offloading ratio

(6)

270

As is appeared in Fig. 5, as the assignment number builds, the normal expense of all the calculation decline, yet the reduction paces of online calculations are marginally greater than disconnected calculations. At the point when the undertaking number is greater than 400, the normal cost of online calculation and disconnected calculations become extremely close. Regarding inclusion speed, disconnected calculation is quicker than the online calculation, they will in general be a steady worth when the task number is 350, while the normal expense of online calculations becomes stable when the undertaking number equivalents to 450. The explanation behind could be that for disconnected calculations, there is a deviation between assessed prompt expense and the real quick cost, it should set aside more effort to update the mistake brought by the estimation.

The numerical outcomes show that the proposed internet learning offloading strategy for portable clients can determine the improved ideal offloading plan to decrease disappointment rate and the client's cost, which is to welcome some ongoing and area mindful applications, for example, video administrations which require a decent client experience. It is especially fit to the situation where the client has to continue moving for quite a while until the strategy can improve itself during the client's development. Be that as it may, during the preparation procedure, it is difficult to ensure that the measures of various sorts of tests are equivalent and this will influence the exhibition of offloading controller. For various clients and various situations, it requests certain expenses and needs clients to move for a while. In our future work, strategies for example, warm beginning of fortification learning or move learning can be embraced in single client's offloading situation to accelerate the learning procedure.

VII. Discussion and Conclusion

This paper proposes a novel plan for QoE-driven calculation offloading issue in a client driven Edge Computing framework. So as to limit the weighted total of the dormancy and vitality utilization, we together enhance the offloading choice, correspondence asset and calculation asset distributions. We have figured the issue as a blend whole number non-straight programming (MINLP) issue, and tackle the enhancement issue through Nature roused calculation. Reproduction results show that our proposed offloading technique can accomplish an improved exhibition on both inactivity and vitality utilization, just as the QoE of end clients.

VIII. References

[1]. M. Armbrust et al., “A view of cloud computing,” Commun. ACM, vol. 53, no. 4, pp. 50–58, 2010.

[2]. S. Ghemawat, H. Gobioff, and S.-T. Leung, “The Google file system,” ACM SIGOPS Oper. Syst. Rev., vol. 37, no. 5, pp. 29–43, 2003.

[3]. J. Dean and S. Ghemawat, “MapReduce: Simplified data processing on large clusters,” Commun. ACM, vol. 51, no. 1, pp. 107–113, 2008.

[4]. K. Shvachko, H. Kuang, S. Radia, and R. Chansler, “The hadoop distributed file system,” in Proc. IEEE 26th Symp.

Mass Storage Syst. Technol. (MSST), Incline Village, NV, USA, 2010, pp. 1–10.

[5]. M. Zaharia, M. Chowdhury, M. J. Franklin, S. Shenker, and I. Stoica, “Spark: Cluster computing with working sets,” in Proc. 2nd USENIX Conf. Hot Topics Cloud Comput., vol. 10. Boston, MA, USA, 2010, p. 10.

[6]. K. Ashton, “That Internet of Things thing,” RFiD J., vol. 22, no. 7, pp. 97–114, 2009.

[7]. H. Sundmaeker, P. Guillemin, P. Friess, and S. Woelfflé, “Vision and challenges for realising the Internet of things,” vol.

20, no. 10, 2010.

[8]. J. Gubbi, R. Buyya, S. Marusic, and M. Palaniswami, “Internet of Things (IoT): A vision, architectural elements, and future directions,” Future Gener. Comput. Syst., vol. 29, no. 7, pp. 1645–1660, 2013.

[9]. “Cisco global cloud index: Forecast and methodology, 2014–2019 white paper,” 2014.

[10]. D. Evans, “The Internet of Things: How the next evolution of the Internet is changing everything,” CISCO White Paper, vol. 1, pp. 1–11, 2011.

[11]. Pooja Chahar, Surjeet Dalal, Deadlock Resolution Techniques: An Overview, International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 pp. 1-5.

[12]. D. Evans, “The Internet of Things: How the next evolution of the Internet is changing everything,” CISCO White Paper, vol. 1, pp. 1–11, 011.

[13]. A. Greenberg, J. Hamilton, D. A. Maltz, and P. Patel, “The cost of a cloud: Research problems in data center networks,”

ACM SIGCOMM Comput. Commun. Rev., vol. 39, no. 1, pp. 68–73, 2008.

[14]. E. Cuervo et al., “MAUI: Making smartphones last longer with code offload,” in Proc. 8th Int. Conf. Mobile Syst. Appl.

Services, San Francisco, CA, USA, 2010, pp. 49–62.

[15]. M. Satyanarayanan, P. Bahl, R. Caceres, and N. Davies, “The case for VM-based cloudlets in mobile computing,” IEEE Pervasive Comput., vol. 8, no. 4, pp. 14–23, Oct./Dec. 2009.

[16]. F. Bonomi, R. Milito, J. Zhu, and S. Addepalli, “Fog computing and its role in the Internet of things,” in Proc. 1st Edition MCC Workshop Mobile Cloud Comput., Helsinki, Finland, 2012, pp. 13–16.

[17]. Boeing 787s to Create Half a Terabyte of Data Per Flight, Says Virgin Atlantic. Accessed on Dec. 7, 2016. [Online].

Available: https://datafloq.com/read/self-driving-carscreate-2-petabytes-data-annually/172

[18]. Self-Driving Cars Will Create 2 Petabytes of Data, What are the Big Data Opportunities for the Car Industry? Accessed on Dec. 7, 2016. [Online]. Available: http://www.computerworlduk.com/news/data/boeing-787screate-half-terabyte-of- data-per-flight-says-virgin-atlantic-3433595/

[19]. Data Never Sleeps 2.0. Accessed on Dec. 7, 2016. [Online]. Available: https://www.domo.com/blog/2014/04/data-never- sleeps-2-0/[19] (2016). OpenFog Architecture Overview. OpenFog Consortium Architecture Working Group. Accessed on Dec. 7, 2016.

[20]. Rani, Uma & Dalal, Surjeet & Kumar, Jugnesh. (2017). Optimizing performance of fuzzy decision support system with multiple parameter dependency for cloud provider evaluation. International Journal of Engineering & Technology. 7. 61.

10.14419/ijet.v7i1.2.8992.

(7)

271

[21]. S. Yi, Z. Hao, Z. Qin, and Q. Li, “Fog computing: Platform and applications,” in Proc. 3rd IEEE Workshop Hot Topics Web Syst. Technol. (HotWeb), Washington, DC, USA, 2015, pp. 73–78.

[22]. K. Ha et al., “Towards wearable cognitive assistance,” in Proc. 12th Annu. Int. Conf. Mobile Syst. Appl. Services, Bretton Woods, NH, USA, 2014, pp. 68–81.

[23]. B.-G. Chun, S. Ihm, P. Maniatis, M. Naik, and A. Patti, “CloneCloud: Elastic execution between mobile device and cloud,” in Proc. 6th Conf. Comput. Syst., Salzburg, Austria, 2011, pp. 301–314.

[24]. A. Rudenko, P. Reiher, G. J. Popek, and G. H. Kuenning, “Saving portable computer battery power through remote process execution,” ACM SIGMOBILE Mobile Comput. Commun. Rev., vol. 2, no. 1, pp. 19–26, 1998.

[25]. G. C. Hunt and M. L. Scott, “The coign automatic distributed partitioning system,” in Proc. OSDI, vol. 99. New Orleans, LA, USA, 1999, pp. 187–200.

[26]. K. Kumar and Y.-H. Lu, “Cloud computing for mobile users: Can offloading computation save energy?” Computer, vol.

43, no. 4, pp. 51–56, Apr. 2010.

[27]. S. Kosta, A. Aucinas, P. Hui, R. Mortier, and X. Zhang, “ThinkAir: Dynamic resource allocation and parallel execution in the cloud for mobile code offloading,” in Proc. IEEE INFOCOM, Orlando, FL, USA, 2012, pp. 945–953.

[28]. L. Zhang et al., “Named data networking (NDN) project,” Xerox Palo Alto Res. Center, Palo Alto, CA, USA, Tech. Rep.

NDN-0001, 2010.

[29]. D. Raychaudhuri, K. Nagaraja, and A. Venkataramani, “MobilityFirst: A robust and trustworthy mobility-centric architecture for the future Internet,” ACM SIGMOBILE Mobile Comput. Commun. Rev., vol. 16, no. 3, pp. 2–13, 2012.

[30]. Dr. Yogesh Kumar Sharma and S. Pradeep (2019), “Deep Learning based Real Time Object Recognition for Security in Air Defense”, “Proceedings of the 13th INDIACom; INDIACom-2019; IEEE Conference ID: 46181 2019 6th International Conference on “Computing for Sustainable Global Development”, 13th - 15th March, 2019 Bharati Vidyapeeth's Institute of Computer Applications and Management (BVICAM), IEEE - New Delhi, (INDIA)”, ISSN No:

0973-7529 , ISBN: 978-93-80544 , Volume : 32, Issue: 8, Pp. 64-67.

[31]. J. Feld, “PROFINET—Scalable factory communication for all applications,” in Proc. IEEE Int. Workshop Factory Commun. Syst., Vienna, Austria, 2004, pp. 33–38.

[32]. J. Cao, L. Ren, W. Shi, and Z. Yu, “A framework for component selection in collaborative sensing application development,” in Proc. 10th IEEE Conf. Coll. Comput. Netw. Appl. Worksharing, Miami, FL, USA, 2014, pp. 104–113.

[33]. F. DaCosta, Rethinking the Internet of Things: A Scalable Approach to Connecting Everything. New York, NY, USA:

ApressOpen, 2013.

[34]. WiFi Network Security Statistics/Graph. Accessed on Dec. 7, 2016. [Online]. Available: http://graphs.net/wifi-stats.html/

[35]. Dr. Yogesh Kumar Sharma and Dr. Surender (2013), “Future Role of Zigbee Technology in Wireless Communication System", Paper published in Grip - The Standard Research International Referred Online Research Journal, ISSN-2278- 8123, Issue No. XVI, Pp. 18-31.

[36]. Open Mhealth Platform. Accessed on Dec. 7, 2016. [Online]. Available: http://www.openmhealth.org/

[37]. K. R. Jackson et al., “Performance analysis of high performance computing applications on the Amazon Web services cloud,” in Proc. IEEE 2nd Int. Conf. Cloud Comput. Technol. Sci. (CloudCom), Indianapolis, IN, USA, 2010, pp. 159–

168.

[38]. A. Li, X. Yang, S. Kandula, and M. Zhang, “CloudCmp: Comparing public cloud providers,” in Proc. 10th ACM SIGCOMM Conf. Internet Meas., 2010, pp. 1–14. [Online]. Available: http://doi.acm.org/10.1145/1879141.1879143 [39]. M. Satyanarayanan, “Mobile computing: The next decade,” SIGMOBILE Mobile Comput. Commun. Rev., vol. 15, no.

2, pp. 2–10, 2011. [Online]. Available: http://doi.acm.org/10.1145/2016598.2016600

[40]. A. Greenberg, J. Hamilton, D. A. Maltz, and P. Patel, “The cost of a cloud: Research problems in data center networks,”

SIGCOMM Comput. Commun. Rev., vol. 39, no. 1, pp. 68–73, Dec. 2008. [Online]. Available:

http://doi.acm.org/10.1145/1496091.1496103

[41]. M. Armbrust et al., “A view of cloud computing,” Commun. ACM, vol. 53, no. 4, pp. 50–58, Apr. 2010. [Online].

Available: http://doi.acm.org/10.1145/1721654.1721672

[42]. A. P. Miettinen and J. K. Nurminen, “Energy efficiency of mobile clients in cloud computing,” in Proc. 2nd USENIX Conf. Hot Topics Cloud Comput., Boston, MA, USA, 2010, p. 4. [Online]. Available:

http://dl.acm.org/citation.cfm?id=1863103.1863107

[43]. N. Ding et al., “Characterizing and modeling the impact of wireless signal strength on smartphone battery drain,”

SIGMETRICS Perform. Eval. Rev., vol. 41, no. 1, pp. 29–40, Jun. 2013. [Online]. Available:

http://doi.acm.org/10.1145/2494232.2466586.