Multi-Method Data Delivery for Green Sensor-Cloud

Green Communications and Computing Multi-Method Data Delivery for Green Sensor-Cloud Chunsheng Zhu, Victor C. M. Leung, Kun Wang, Laurence T. Yang, and Yan Zhang The authors discuss the potential applications and recent work on SC, and observe two issues regarding green SC. Further, motivated by solving these two issues, they propose a multi-method data delivery scheme for SC users. strategically incorporates four kinds of delivery: delivery from cloud to SC users; delivery from WSN to SC users; delivery from SC users to SC users; and delivery from cloudlet to SC users. Abstract Delivering sensory data to users anytime and anywhere if there is network connection, sensor-cloud (SC), which integrates WSNs and cloud computing, is attracting growing interest from both academia and industry. This article discusses the potential applications and recent work about SC and observes two issues regarding green SC. Further, motivated by solving these two issues, this article proposes a Multi-Method Data Delivery () scheme for SC users. strategically incorporates four kinds of delivery: delivery from cloud to SC users; delivery from WSN to SC users; delivery from SC users to SC users; and delivery from cloudlet to SC users. Compared to exclusive data delivery from cloud to SC users, evaluation results show that could achieve lower delivery cost or less delivery time for SC users. Introduction Recently, motivated by incorporating the ubiquitous data gathering ability of wireless sensor networks (WSNs) as well as the powerful data storage and data processing capabilities of cloud computing (CC), sensor-cloud (SC) [1] [2] is receiving growing attention from both the academic and industrial communities. Basically, integrating WSNs and CC, as shown in Fig. 1, the sensory data is gathered by the ubiquitous sensor nodes (e.g., temperature sensor nodes, humidity sensor nodes, motion sensor nodes) in WSNs and transmitted to the powerful data centers in the cloud. Then the sensory data is stored and processed by the cloud and further delivered to SC users on demand. With such integration, from the perspective of users, SC enables them to obtain their required sensory data anytime and anywhere if there is network connection. From the view of WSNs and the cloud, SC complements them. For example, in the cloud, the service the cloud offers can be greatly enriched by providing the services (e.g., environmental monitoring, healthcare monitoring, landslide detection, forest fire detection) that WSN provides [3]. Regarding WSNs, the utility of WSNs could be enhanced, by being able to serve multiple applications via the cloud. Generally, in such integration, there are three main entities: a sensor network provider (SNP), which enables the WSN; the cloud service provider (CSP), which offers the cloud, and the SC user. Discussing the potential applications and recent work regarding SC, this article observes two issues concerning green SC. Then, triggered by solving these two issues, this article proposes a Multi-Method Data Delivery () mechanism for SC users. Particularly, strategically combines four kinds of delivery: delivery from the cloud to SC users; delivery from the WSN to SC users; delivery from SC users to SC users; and delivery from a cloudlet to SC users. In contrast to exclusive data delivery () from the cloud to SC users, evaluation results show that could obtain lower delivery loss or less delivery time for SC users. For the rest of this article, the potential applications of SC are presented in the next section. Following that we review the recent work on SC and present the two issues regarding green SC. The proposed scheme is then introduced. We next perform the evaluation with respect to and. This article is concluded in the final section. Potential Applications of SC SC has a lot of exciting potential applications [4]. For instance, concerning real-time agriculture monitoring, WSNs comprise a variety of sensor nodes (e.g., soil moisture sensor nodes, air sensor nodes, temperature sensor nodes, CO 2 concentration sensor nodes, and camera sensor nodes) that can be arranged to collect various information about the crops on a farm. These data can be further analyzed by the cloud in real time in order to track the health of the crops. With respect to real-time transportation monitoring, WSNs that include different kinds of sensor nodes (e.g., pressure sensor nodes, image sensor nodes, video sensor nodes, and alcohol gas sensor nodes) can be used for gathering vehicle and driver information. After the cloud incorporates the collected information in real time, the level of fuel and the vehicle arrival time as well as the status of the driver can be tracked, predicted, and observed, respectively. Regarding real-time tunnel monitoring, WSNs including light sensor nodes can be utilized to sense the light levels inside a tunnel. Meanwhile, the cloud can analyze the sensed light levels in real time so that the light intensity can be automatically adjusted to save energy spent unnecessarily for lighting throughout the day. About real-time wildlife monitoring, WSNs consisting of various types of sensor nodes (e.g., video sensor nodes) can be deployed into a wide Digital Object Identifier: 1.119/MCOM.217.16822 Chunsheng Zhu and Victor C. M. Leung are with The University of British Columbia; Kun Wang is with Nanjing University of Posts and Telecommunications; Laurence T. Yang is with St. Francis Xavier University. Yan Zhang is with the University of Oslo, 373 Oslo. 176 163-684/17/$25. 217 IEEE IEEE Communications Magazine May 217

Wireless sensor networks (WSNs) Temperature sensor Humidity sensor Figure 1. An instance of SC. Video sensor Motion sensor Data queries Data transmission field, collecting information about the wildlife sanctuaries, activities, and so on. With the powerful cloud, which stores and processes the gathered information, real-time monitoring and further protecting the wildlife (e.g., endangered species) can be achieved. Recent Work about SC Focusing on cloud-based WSNs, the aim of [5] is enhancing the lifetime of the WSN integrated with the cloud. Specifically, two collaborative location-based sleep scheduling approaches are introduced for WSNs. The strategy is to dynamically determine the awake or asleep state of each sensor node to decrease energy consumption of the integrated WSN, considering the locations of mobile cloud users. With respect to the WSN-based cloud, the purpose of [6] is reducing the expected completion time for the CC integrated with WSN. Particularly, two job scheduling schemes (i.e., priority-based two-phase Min-Min and priority-based two-phase Max-Min) are described for CC. The technique executes WSN related cloud tasks in phase 1, while executing other ordinary cloud tasks in phase 2. Concerning SC integration, a sensory data processing framework is shown in [7], aimed at transmitting desirable sensory data to the mobile cloud users in a fast, reliable, and secure manner. The mechanism incorporates the WSN gateway, the cloud gateway, the cloud, and the mobile users to perform various functions (e.g., data traffic monitoring, data filtering, data prediction, data recommendation, data compression, data decompression, data security). Another sensory data delivery scheme is presented in [8], toward offering more useful data reliably to the mobile cloud from a WSN. The idea is making the WSN gateway selectively transmit the sensory data to the cloud based on the time and priority features of the data requested by the mobile user, while utilizing the priority-based sleep scheduling to save the energy consumption of the WSN. An authenticated trust and reputation calculation and management system is exhibited in [9], targeted at helping a cloud service user (CSU) choose a desirable CSP and assisting the CSP in selecting an appropriate SNP. The scheme incorporates the authenticity of the CSP and SNP; the attribute requirement of the CSU and CSP; and the cost, trust, and reputation of the service of the CSP and SNP. A trust-assisted SC is designed in [1], devoting effort to improve the quality of service (QoS) Cloud Data center 1 Users User 1 Send data requests User 2 Data center 2 Data center 3 Reply data requests User 3 Data server of the sensory data experienced by SC users. The method utilizes trusted sensors (i.e., sensors with trust values surpassing a threshold) in the WSN for gathering and transmitting sensory data, while using trusted data centers (i.e., data centers with trust values surpassing a threshold) in the cloud for storing, processing, and delivering the sensory data to users. Five pricing models are devised in [11], induced by offering guidance for future research regarding SC pricing. The pricing designs consider the following factors: the lease period of the SC user; the required working time of SC; the SC resources utilized by the SC user; the volume of sensory data obtained by the SC user; and the SC path that transmits the sensory data from the WSN to the SC user, respectively. To the best of our knowledge, the sensory data delivery of SC in all the above work is from the cloud to SC users. With such delivery, the following two issues regarding green SC could exist. Issue 1: Since SC users might request the same data from the cloud, the cloud might deliver large amounts of the same data to SC users. A large number of repeated data transmissions from the cloud to SC users exclusively increases the demand regarding the energy and resources as well as the bandwidth of SC. Issue 2: When multiple SC users request data from the cloud simultaneously, a lot of data needs to be delivered from the cloud to multiple SC users at the same time. Substantial data delivery from the cloud to multiple SC users exclusively also increases the requirement with respect to the energy and resources as well as the bandwidth of SC. In both of the above cases, we can observe that the delivery cost (e.g., utilized energy and resources as well as bandwidth) for providing data to SC users is increased. Furthermore, in terms of an SC with certain energy and resources as well as bandwidth, the delivery time for offering data to SC users is also increased. The Proposed scheme Overview Motivated by solving the above observed two issues, the scheme is proposed. Particularly, as shown in Fig. 2, the following four methods are incorporated by for delivering sensory data to SC users: 1. 1 (delivery from cloud to SC users) 2. 2 (delivery from WSN to SC users) Two collaborative location-based sleep scheduling approaches are introduced for WSNs. The strategy is dynamically determining the awake or asleep state of each sensor node to decrease energy consumption of the integrated WSN, considering the locations of mobile cloud users. IEEE Communications Magazine May 217 177

In terms of a certain delivery method in, the delivery cost and delivery time probably are also various in different situations. Thus, the appropriate delivery method (s) in different conditions need (s) to be used, to better satisfy the SC user s requirement regarding delivery cost or delivery time. Delivery scheme Communication method 1 (delivery from cloud to SC users) Network communication 2 (delivery from WSN to SC users) Base station communication 3 (delivery from SC to SC users) Device-to-device communication 4 (delivery from cloudlet to SC users) Local area network communication User Wireless sensor networks (WSNs) Data queries Data transmission 2 1 Cloud Data center 1 4 3 Cloudlet Send data requests Users User 1 User 2 Data center 2 Data center 3 Reply data requests User 3 Temperature sensor Humidity sensor Video sensor Motion sensor Data server Figure 2. An instance of. 3. 3 (delivery from SC users to SC users) 4. 4 (delivery from cloudlet to SC users) Regarding item 1, 1 (delivery from cloud to SC users), the sensory data is delivered via the network communication, for example, wideband code-division multiple access (WCDMA), LTE, or WiMAX. About item 2, 2 (delivery from WSN to SC users), the sensory data is delivered via base station communication. With respect to item 3, 3 (delivery from SC users to SC users), the sensory data is delivered via device-to-device communication. For item 4, 4 (delivery from cloudlet to SC users), the sensory data is delivered via the local area network communication. Here, available for use by nearby mobile devices, a cloudlet [12] is a resource-rich and trusted computer or cluster of computers well connected to the Internet. Delivery Rules The delivery rules consider the following elements: The delivery methods that are available for the SC user might be different when the SC user is in various locations. The sensory data requested by the SC user needs to be delivered through the delivery method(s). For each delivery method in, there is an associated delivery cost and delivery time. Particularly, regarding different delivery methods in, the delivery cost and delivery time probably vary in different conditions. In terms of a certain delivery method in, the delivery cost and delivery time probably also vary in different situations. Thus, the appropriate delivery method(s) in different conditions need(s) to be used, to better satisfy the SC user s requirement regarding delivery cost or delivery time. Specifically, the rules to utilize the delivery method(s) are shown as follows. Rule 1: Based on the location of the SC user, the delivery method(s) available to the SC user is (are) determined by the SC. Rule 2: Taking into account the data required by the SC user, the delivery method(s) that can deliver the needed sensory data is (are) decided by the SC. Rule 3: Considering the delivery cost and delivery time of the delivery method(s), the specific delivery method(s) is (are) utilized by the SC based on the service level agreement of the SC user. Regarding rule 1, the locations of SC users are obtained by SC through a mobile application that dynamically uploads the location of the SC user to SC [13]. About rule 2, if the sensory data required by the SC user needs to be delivered by multiple delivery methods, multiple delivery methods become the candidates. Otherwise, only one delivery method is used. With respect to rule 3, if the SC user wants to obtain the data with minimum delivery time, the delivery method(s) with the minimum delivery time is (are) utilized. If the SC user wants to achieve the data with a minimum delivery cost, the delivery method(s) with the minimum delivery cost is (are) used. Otherwise, the delivery method(s) with a satisfactory delivery time (i.e., the delivery time is less than a threshold) and a minimum delivery cost is (are) utilized, or the delivery method(s) with a satisfactory delivery cost (i.e., the delivery cost is less than a threshold) and a minimum delivery time is (are) used. Delivery and Application Analysis The purpose of is to better cater for the SC user s delivery requirement, since the delivery cost or delivery time with might not satisfy the SC user s delivery requirement in some cases. For example, for issues 1 and 2 discussed above, in the case that SC users request the same data from the cloud or multiple SC users request the data from the cloud simultaneously, the data could be offered to the SC users by intelligently utilizing different delivery method (s) in, instead of from the cloud to SC users. Since 178 IEEE Communications Magazine May 217

Wireless sensor networks (WSNs) Temperature sensor Humidity sensor Video sensor Motion sensor Figure 3. application in SSC. Social networks (SNs) User 2 1 Cloud 4 3 Cloudlet Users Data center 1 User 1 Data queries Data transmission Data center 2 Data center 3 Send data requests Reply data requests User 2 User 3 Data server The aim of is to strategically incorporate the four delivery methods to generate lower delivery cost or less delivery time for the SC user, while offering the needed data to the SC user. As a result, different (s) might be utilized for delivering the sensory data to the SC user in different conditions. different delivery methods in have varying delivery cost and delivery time in different conditions, the delivery cost for providing data to SC users with might be lower than that with. Similarly, the delivery time for offering data to SC users with might be less than that with. In other words, the aim of is to strategically incorporate the four delivery methods (i.e., 1, 2, 3, and 4) to generate lower delivery cost or less delivery time for the SC user, while offering the needed data to the SC user. As a result, different s might be utilized for delivering the sensory data to the SC user in different conditions. For instance, 4 might be used when the SC user is in a coffee shop in which a cloudlet could offer the sensory data, while 3 probably could be used when the SC user is in a classroom where other SC users could provide the sensory data. 2 might be utilized when the SC user is very close to the WSN offering the sensory data, while 1 probably could be utilized when the SC user is very far from home and the cloud could provide the sensory data. Moreover, multiple delivery methods in could be used if one delivery method (e.g., 4) is not sufficient to offer all the sensory data that the SC user requests. Furthermore, regarding, it could also be utilized in the future social sensor cloud (SSC) [14]. As shown in Fig. 3 about SSC, the SC users form social networks (SNs) [15], which connect and complement the WSNs, cloud, and cloudlet. In particular, SNs are networks formed by social members (e.g., individuals, organizations) who interact with each other. With SNs, the resources and services of the WSNs, cloud, and cloudlet could be shared, leveraging the relationships established between members of an SN. In such a manner, the resources and services requested by the SC users could be substantially reduced. Then the delivery cost and delivery time with might be further decreased. With further decreased delivery cost or delivery time, the SC user s delivery requirement could be even better satisfied. Evaluation Evaluation Setup To evaluate the delivery time and delivery cost of and for SC, the following two evaluations are performed as case studies. In both evaluations, there is an SC consisting of a WSN, a cloud, a cloudlet, and a number of SC users. In terms of, the sensory data is delivered to the SC users from the cloud exclusively. For, the sensory data is delivered to the SC users cooperatively with 1, 2, 3, and 4. For evaluation 1, it is assumed that the delivery times to an SC user with 1, 2, 3, and 4 are.4 s,.3 s,.2 s, and.1 s, respectively. The delivery cost to an SC user with 1, 2, 3, and 4 are US$.8, US$.6, US$.4, and US$.2, respectively. Evaluation 1 is conducted in the two scenarios below. Scenario 1: 1, 2, 3, and 4 are each used to serve 25 percent of SC users, respectively. The number of SC users ranges from to (increased by units of ). This scenario is for evaluating the impacts of the number of SC users on SC s delivery time and SC s delivery cost with and. Scenario 2: There are 5 SC users. The percentage of SC users 1 serves ranges from 1 to 9 percent (increased by units of 1 percent). 2, 3, and 4 are utilized to serve the remaining users equally. This scenario is to evaluate the impacts of the utilization rate of 1 on SC s delivery time and SC s delivery cost with and. Concerning evaluation 2, there are 5 SC users, served by 1, 2, 3, and 4 equally. Moreover, it is given that the delivery times to an SC user with 3 and 4 are.4 s and.2 s, respectively. The delivery cost to an SC user with 3 and 4 are US$.8 and US$.4, respectively. Evaluation 2 is implemented in the following two scenarios. Scenario 3: The delivery time and delivery cost to an SC user with 1 are.8 ss and IEEE Communications Magazine May 217 179

5 45 9 4 8 35 7 3 25 6 5 4 15 3 5 25 3 4 5 6 7 8 9 Number of SC users (a) 5 3 4 5 6 7 8 9 Number of SC users (b) 225 45 4 175 35 15 125 3 25 75 15 5 25 5.1.2.3.4.5.6.7.8.9 Utilization rate of 1 (c).1.2.3.4.5.6.7.8.9 Utilization rate of 1 (d) Figure 4. Evaluation 1: comparison of a) delivery time and b) delivery cost b) with and in scenario 1; comparison of c) delivery time and d) delivery cost with and in scenario 2. US$1.6, respectively. The delivery time to an SC user with 2 ranges from.6 to 1.4 s (increased in units of.1 s). The delivery cost to an SC user with 2 ranges from US$1.2 to US$2., increased by units of US$.1). This scenario is for evaluating the impact of 2 s delivery time on SC s delivery time and the impact of 2 s delivery cost on SC s delivery cost with and. Scenario 4: The delivery time and delivery cost to an SC user with 2 are.6 s and US$1.2, respectively. The delivery time to an SC user with 1 ranges from.8 to 1.6 s (increased by units of.1 s). The delivery cost to an SC user with 1 ranges from US$1.6 to US$2.4 (increased by units of US$.1). This scenario is to evaluate the impact of 1 s delivery time on SC s delivery time and the impact of 1 s delivery cost on SC s delivery cost with and. Evaluation Results Figures 4 and 5 present the evaluation 1 results and evaluation 2 results about the delivery time and delivery cost with and for SC in the two different scenarios, respectively. As shown in these figures, it can be clearly obtained that the delivery time or delivery cost of is better than that of 18 IEEE Communications Magazine May 217

5 45 9 4 8 35 7 3 25 6 5 4 15 3 5.6.7.8.9 1 1.1 1.2 1.3 1.4 Delivery time of 2 (s) 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Delivery time of 2 (s) (a) 15 (b) 9 135 8 1 7 15 6 5 4 9 75 6 3 45 3 15.8.9 1 1.1 1.2 1.3 1.4 1.5 1.6 Delivery time of 1 (s) (c) 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 Delivery time of 1 ($) (d) Figure 5. Evaluation 2: comparison of a) delivery time and b) delivery cost with and in scenario 3; comparison of c) delivery time and d) delivery cost with and in scenario 4. in terms of the above case studies. Particularly, from Figs. 4a and 4c and 5a and 5c, it can be observed that the SC s delivery times with are always less than those with in scenarios 1, 2, 3, and 4. In addition, based on Figs. 4b and 4d and 5b and 5d, it can be perceived that the SC s delivery costs with are also always lower than those with in all four scenarios. Conclusion Attracting growing interest from both the academic and industrial communities by integrating WSNs and CC, SC delivers sensory data to users anytime and anywhere if there is network connection. In this article, the potential applications and recent work with respect to SC have been discussed, and two research issues about green SC have been identified. Further, the scheme has been proposed, induced by solving the observed two research issues. Specifically, the following four methods that deliver sensory data to SC users are strategically incorporated in : Delivery from cloud to SC users Delivery from WSN to SC users Delivery from SC users to SC users Delivery from cloudlet to SC users Evaluation results have also been presented about IEEE Communications Magazine May 217 181

With SNs, the resources and services of the WSNs, cloud and cloudlet could be shared, leveraging the relationships established between members of a SN. In such a manner, the resources and services requested by the SC users could be substantially reduced. Then the delivery cost and delivery time with might be further decreased. and, demonstrating that could achieve lower delivery cost or less delivery time for SC users. Acknowledgment This work was partially supported by a Four Year Doctoral Fellowship from The University of British Columbia and funding from the Natural Sciences and Engineering Research Council of Canada, the ICICS/TELUS People & Planet Friendly Home Initiative at The University of British Columbia, TELUS and other industry partners. This work was partially supported by the National Natural Science Foundation of China (NSFC) under Grant 61572262, the NSF of Jiangsu Province under Grant BK2141427. This work was partially supported by the project IoTSec - Security in IoT for Smart Grids, with number 248113/O7 part of the IKTPLUSS program funded by the Norwegian Research Council. This research is partially supported by the projects 2479/F2 funded by the Research Council of Norway. References [1] C. Zhu et al., Sensor-Cloud and Power Line Communication: Recent Developments and Integration, Proc. 14th IEEE Int l. Conf. Depend., Autonomic, Secure Comp., 216, pp. 32 8. [2] C. Zhu et al., Towards Integration of Wireless Sensor Networks and Cloud Computing, Proc. 7th IEEE Int l. Conf. Cloud Comp. Tech. Sci., 215, pp. 491 94. [3] C. Zhu et al., Insights of Top-K Query in Duty-Cycled Wireless Sensor Networks, IEEE Trans. Ind. Electron., vol. 62, no. 2, Feb. 215, pp. 1317 28. [4] A. Alamri et al., A Survey on Sensor-Cloud: Architecture, Applications, and Approaches, Int l. J. Distrib. Sensor Net., vol. 9, no. 2, Feb. 213, pp. 1 18. [5] C. Zhu et al., Collaborative Locationbased Sleep Scheduling for Wireless Sensor Networks Integrated with Mobile Cloud Computing, IEEE Trans. Comp., vol. 64, no. 7, July 215, pp. 1844 56. [6] C. Zhu et al., Job Scheduling for Cloud Computing Integrated with Wireless Sensor Network, Proc. 6th IEEE Int l. Conf. Cloud Comp. Tech. Sci., 214, pp. 62 69. [7] C. Zhu et al., A Novel Sensory Data Processing Framework to Integrate Sensor Networks with Mobile Cloud, IEEE Sys. J., vol. 1, no. 3, Sept. 216, pp. 1125 36. [8] C. Zhu et al., Towards Offering More Useful Data Reliably to Mobile Cloud from Wireless Sensor Network, IEEE Trans. Emerging Topics Comp., vol. 3, no. 1, Mar. 215, pp. 84 94. [9] C. Zhu et al., An Authenticated Trust and Reputation Calculation and Management System for Cloud and Sensor Networks Integration, IEEE Trans. Info. Forensics Security, vol. 1, no. 1, Jan. 215, pp. 118 31. [1] C. Zhu et al., Trust Assistance in Sensor-Cloud, Proc. IEEE Conf. Comp. Commun. Wksps., 215, pp. 342 47. [11] C. Zhu et al., Pricing Models for Sensor-Cloud, Proc. 7th IEEE Int l. Conf. Cloud Comp. Tech. Sci., 215, pp. 454 57. [12] M. Satyanarayanan et al., The Case For VM-Based Cloudlets in Mobile Computing, IEEE Pervasive Comp., vol. 8, no. 4, Oct.-Dec. 9, pp. 14 23. [13] G. Ananthanarayanan et al., Startrack: A Framework for Enabling Track-Based Applications, Proc. 7th Int l. Conf. Mob. Sys., Appl., Serv., 9, pp. 27 2. [14] C. Zhu et al., Green Internet of Things for Smart World, IEEE Access, vol. 3, Nov. 215, pp. 2151 62. [15] Y. Jiang and J. C. Jiang, Understanding Social Networks from a Multiagent Perspective, IEEE Trans. Parallel Distrib. Sys., vol. 25, no. 1, Oct. 214, pp. 2743 59. Biographies Chunsheng Zhu is a postdoctoral research fellow in the Department of Electrical and Computer Engineering, University of British Columbia, Canada. He received his Ph.D. degree in electrical and computer engineering from the University of British Columbia in 216. His current research interests mainly include wireless sensor networks, cloud computing, Internet of Things, social networks, and security. Victor C. M. Leung [S 75, M 89, SM 97, F 3] is a professor in the Department of Electrical and Computer Engineering and holder of the TELUS Mobility Research Chair, University of British Columbia. His research is in the areas of wireless networks and mobile systems. He is a Fellow of the Royal Society of Canada, a Fellow of the Canadian Academy of Engineering, and a Fellow of the Engineering Institute of Canada. Kun Wang [M 13] is an associate professor in the School of Internet of Things, Nanjing University of Posts and Telecommunications, China. He received his Ph.D. degree from the School of Computing, Nanjing University of Posts and Telecommunications, in 9. In 216, he was a research fellow with the School of Computer Science and Engineering, University of Aizu, Fukushima, Japan. His current research interests include information security, ubiquitous computing, and wireless communications technologies. Laurence T. Yang [M 97, SM 15] is a professor in the Department of Computer Science, St. Francis Xavier University, Canada. His research interests include parallel and distributed computing, embedded and ubiquitous/pervasive computing, and big data. He has published more than 22 papers in various refereed journals (around 4 percent in top IEEE/ACM transactions and journals). His research has been supported by the National Sciences and Engineering Research Council, and the Canada Foundation for Innovation. Yan Zhang [SM 1] is a professor in the Department of Informatics, University of Oslo, Norway. He is also a chief research scientist at Simula Research Laboratory, Norway. He is an Associate Technical Editor of IEEE Communications Magazine, an Editor of IEEE Transactions on Green Communications and Networking, and an Editor of IEEE Communications Surveys & Tutorials. His current research interests include next-generation wireless networks leading to 5G, green, and secure cyber-physical systems. 182 IEEE Communications Magazine May 217