Ahmad 2018

Accepted Manuscript Socio-cyber network: The potential of cyber-physical system to define human behaviors using Big Data analytics Awais Ahmad, Muhammad Babar, Sadia Din, Shehzad Khalid, Muhammad Mazhar Ullah, Anand Paul, Alavalapati Goutham Reddy, Nasro Min-Allah

PII: DOI: Reference:

S0167-739X(17)30778-1 https://doi.org/10.1016/j.future.2017.12.027 FUTURE 3865

To appear in:

Future Generation Computer Systems

Received date : 26 April 2017 Revised date : 27 November 2017 Accepted date : 22 December 2017 Please cite this article as: A. Ahmad, M. Babar, S. Din, S. Khalid, M.M. Ullah, A. Paul, A.G. Reddy, N. Min-Allah, Socio-cyber network: The potential of cyber-physical system to define human behaviors using Big Data analytics, Future Generation Computer Systems (2018), https://doi.org/10.1016/j.future.2017.12.027 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Socio-Cyber Network: The Potential of Cyber-Physical System to Define Human Behaviors using Big Data Analytics Awais Ahmad1, Muhammad Babar2, Sadia Din3, Shehzad Khalid4, Muhammad Mazhar Ullah3, Anand Paul3, Alavalapati Goutham Reddy4*, Nasro Min-Allah5 1

Department of Information and Communication Engineering, Yeungnam University, Gyeongsan, Republic of Korea Department of Computer Software Engineering, National University of Sciences and Technology, Islamabad, Pakistan 3 School of Computer Science and Engineering, Kyungpook National University, Daegu, Korea 4 Department of Computer & Information Security, Sejong University, Seoul, 05006, South Korea 5 College of Computer Science and Information Technology, University of Dammam, KSA 2

[email protected], [email protected], [email protected], [email protected], [email protected], [email protected], *[email protected], [email protected]

Abstract— The growing gap between users and the Big Data analytics requires innovative tools that address the challenges faced by big data volume, variety, and velocity. Therefore, it becomes computationally inefficient to analyze such massive volume of data. Moreover, advancements in the field of Big Data application and data science leads toward a new paradigm of human behavior, where various smart devices integrate with each other and establish a relationship. However, majority of the systems are either memoryless or computational inefficient, which are unable to define or predict human behavior. Therefore, keeping in view the aforementioned needs, there is a requirement for a system that can efficiently analyze a stream of Big Data within their requirements. Hence, this paper presents a system architecture that integrates social network with the technical network. We derive a novel notion of ‘Socio-Cyber Network’, where a friendship is made based on the geo-location information of the user, where trust index is used based on graphs theory. The proposed graph theory provides a better understanding of extraction knowledge from the data and finding relationship between different users. To check the efficiency of the proposed algorithms exploited in the proposed system architecture, we have implemented our proposed system using Hadoop and MapReduce. MapReduce for cyber-physical system (CPS) is supported by a parallel algorithm that efficiently process a huge volume of data sets. The system is implemented using Spark GraphX tool at the top of the Hadoop parallel nodes to generate and process graphs with near real-time. Moreover, the system is evaluated in terms of efficiency by considering the system throughput and processing time. The results show that the proposed system is more scalable and efficient. Index Terms— Big Data, socio-cyber network, human behavior, graphs, friendship, trust index. I.

INTRODUCTION

Recently, the scope of the digital world in increased in a way that a number of smart objects are connecting in a huge quantity. It is due to the exploitation of low-cost internet extensions, air-interface, and revolution in miniaturization. Hence, it results in decreasing the cost of a smart device. These smart devices are referred to physical devices, which has the capability to sense physical stimulate, aggregate the data, and

interact with other physical objects. Moreover, up to certain extent, and due to the terminology of ‘smartness’, these objects are intelligent in nature that show intellectual behavior in establishing connections and interaction with other physical objects based on some protocols, and operate autonomously. Some of the general examples are building automation system, smart watches, transportation system, intelligent healthcare system, smart watch, smart TV, security system, etc. [1-2]. Such interaction and operation are increased when they are connected via the internet, which makes the terminology of ‘cyber-physical system (CPS)’. Thus, improving the quality of life, performance and security. So far, CPS plays a vital role in supporting interactions between various objects so that they can send their data, to carry out actions, and provide a bridge between heterogeneous objects to interact among themselves. Thus, for this purpose, the notion of a web of things and service oriented architectures is also introduced that supports heterogeneity [3-4]. Having understood the fact that CPS plays a vital role in providing a platform through which devices are interconnected. Now, a challenge is how humans and smart objects will interact with each other? Initially, the role of cyber was providing a human-to-human connection in which the basic content of communication was designed by humans, i.e., to be consumed by other humans. However, with the recent advent in CPS and web of things terminologies, the designing of contents are changed from human to devices. Thus, it has an overwhelming influence upon our lives, i.e., these days, a secret and non-touchable research area. And also, it requires an understanding of how CPS plays a key role in enhancing and understanding our smart environment. Moreover, the enhancement of CPS starting its impact in the social networks. For instance, “Like-Art” [5], in which the perspective of social network and “Like it” from Facebook is pooled with the physical contact over undulating hand. Such activity is read by sensors that are connected to the Internet. Therefore, by doing so, visitors are then able to specify that an explicit representation is favorite in the cyber world by interacting with the physical environment.

The exploitation of the social network with the internet is investigated in [6], where the human social network is used to share various resources presented by smart devices. In the similar context, the establishment of Ubiquitous Internet of Things architecture stimulated by the societal association of human being [7]. Moreover, the bond between social and technical network is discussed in [8], where they came up with a terminology of ‘socio-technical networks’. This concept is widely used in the context of Internet of Things. Having such information where social network integrates with the technical network tends to connect millions of smart devices [2], where users can understand various activities, such as likes and dislikes. Other activities are, e.g., which places are visited often, and healthcare system using wireless body area sensor network, in which devices interact with each other using the Internet. Due to such interaction of smart devices, a rapid growth is seen in generating a massive volume of data. This massive amount of data requires some sort of medium where it can be stored, e.g., databases. This growth is in the form of three Vs. Such a volume, velocity, and veracity. Moreover, the massive growth uses the common medium for communication, i.e., the Internet. According to the report published by CISCO, humans that use smart devices generating 2.5 quintillion byte of data in each day [9]. Additionally, it is also stated that relational databases run over archived data in response to postmortems in the case of credit card transactions. In addition, various data processing techniques are also developed to maintain data aggregation in a fault-tolerant manner. Such data is access by means of any query mechanism generated by users. In such circumstances, generic big data analytics are achieved by traditional database vendors. These are used for data aggregation. These vendors are generally software based, or they only provide analytical services, which runs over third party services. Generally, traditional databases are unable to predict human behavior. Nowadays, interaction models are designed that are based on human inquiries about the information given by objects (i.e., human-object interaction). However, this model will be shifted to object-object interaction in a near future. Therefore, in this way, the object will look for other so that to provide compound facilities that will be useful for humans. This mechanism may increase the interaction complexity. Furthermore, various challenges can be found, in which scalability issues are one of the keys confront in said case [11]. In such scenario, several techniques are given for real-time analysis of human behavior [11-12]. However, the majority of these schemes are unable to process and analyze big data. Thus, based on the aforementioned explanation and challenges, there is a need to change the existing traditional paradigms of analyzing big data to predict and assist human behavior. To do so, a system architecture requires that is based on the interaction of various activities of users, let say friendship selection. Moreover, advancement in the field of big data analytics is one of the utmost crucial factors to analyze human behavior. Therefore, this paper we extend the concept of smartbuddy [41] presents system architecture to analyze human behavior based on knowledge extraction. Knowledge extract plays a vital role in providing the depth of experience, which helps the user in creating reliability index. Moreover, trust index is also examined by exploiting graph based search using

the social network, which finally helps in analyzing the behavior of human by considering friends management, social monitoring, and social analysis. The proposed system architecture is tested and implemented using Hadoop ecosystem by considering various datasets, such as social network, healthcare, and intelligent transportation system. II.

RESEARCH CONTRIBUTION

The main contribution of this paper is as follows.  At first, we present a hierarchical framework for integrating social network with the technical network. The framework first internments the scalable features of CPS, which helps in the extension of network commodities. Moreover, the proposed system architecture assists the network in data aggregation. The exploitation of reliability index, trust index, and reputation index in the architecture bridges the social user activities of the users.  Moreover, we proposed a novel notion of ‘socio-cyber network’, which is based on graph analysis that connects edges of various attributes involved in the social and technical network. The proposed graph-based technique facilitates the system architecture to provide in-depth knowledge of the big data being generated by smart device connected to each other.  Finally, we conduct the extensive simulation using Hadoop-based analysis by considering the simulated scenario of the social and technical network. From the results and discussion, we show that the proposed system architecture is feasible for analyzing human behavior using big data. The remainder of this paper is organized as follows. In Section 3, we give a detailed description of the background and related studies. In Section IV, we presents the basic level knowledge about the social structure of human behavior. In Section v, we propose multilevel system design architecture for communication. Also, we described proposed system architecture and their parts in the same section. In Section VI, a detailed simulation results and discussion are provided. Finally, Section 5 offers the conclusion of the paper. III. RELATED WORK MapReduce programming paradigm producing a large amount of datasets that is responsible for the extensive diversity of real world tasks [13]. MapReduce divides input data into small independent chunks which dealt in a parallel manner completely. The MapReduce architecture classifies the maps outputs and sends to the reduce job. Basically, the input and output of the task are kept in the file system. MapReduce is a parallel programming model, which perform three main task at the same time i.e., simplicity, load balancing, and fault tolerance. The Google File System (GFS) normally inspired from the MapReduce model gives the reliability and efficiency of data storage required for large databases applications [14]. MapReduce model motivated by functional languages. Functional languages have a map and reduce primeval exist in functional languages. Depending on the framework

requirement numerous execution can be feasible in MapReduce platform. Few recently executions are existed in literature work e.g. networked machines clustering [13], shared memory multi-core system techniques [15-16], graphic processors and asymmetric multi-core processors approach [17]. The majority of exploratory research depend on the novel technologies this is disclosed by current research. Now people are now totally depending on these emergency technologies. So these smart things are very perspicacious and capable of interfacing with people. In the literature, authors stated about Things. These things are interconnected and communicate with each other with the help of internet [18]. This is labeled as neologism Blobject i.e., objects that blog. Furthermore, Embodied Microblogging (EM) 15 has many issues in IoT. The focal point of the author in this study is human to human instead of human to things. According to a survey, people think that IoT and social network are related to each other. To be very careful they assumed internet vision as ‘Ubiquitous IoT’ [19]. Ubiquitous IoT is almost near to the notion of social institute model. It seems that this conviction does not understand the fundamental keynote of the social network in IoT. Miscellaneous platforms and structural designs are designed for illustrating the combination of social network and IoT in the last decade [2] [20-21]. This type of combination helps to design the social organization network model in IoT. The function of the web cannot be undervalued in which, with the help of the web various devices are interconnected with each other [3]. This idea is named as Web of Things (WoT). A three-tier architecture is designed to access the system without difficulty through uniform web services RESTfull APIs [22]. This proposed architecture interpret raw information into ordinary language and after that translate ordinary language into machine readable language. Paraimpu is a social tool used to comprise and easily interconnect smart things, mashup services, and devices to make personal applications in Internet of Things. It also connects the Web to physical Things, and APIs to create personalized applications. The proposed scheme exploits the social network for user verification, import users accounts, friend list and communication with each other. Likewise, a social access controller (SAC) is designed to spread the social network idea [23]. This proposed system verifies proxy between smart objects and users. By using social network verification API, SAC assist in handling access control depend on the current social organization of a network. Moreover, various other schemes of the social network are designed for many different demands. For example, users can share data with each other with the help of SenseShare application [24]. But the user cannot communicate with sensors using SenseShare. Additionally, Friends and Things applications (FAT) assists the user to share their data with trusted users. In the literature review, there are many types of research done to handle a huge amount of data and provide IoT services for a wide scope of frameworks. Although big data has numerous difficulties but IoT begin to be a very valuable study in research [25] [43], and [44]. IoT uses many big data tools that are very important for analyzing big data for instance MapReduce, Cassandra, NoSQL, etc. Furthermore, in the study different proposed architecture are used for implementation of

big data and handling with online and offline huge data which comes from IoT. Data which comes from connected things can be store and analyze with the help of different storage facilities such as cloud and big data techniques [26], [40], and [42]. These storage services enhance data scalability, flexibility, availability, and adaptability. Beside that when connecting social network with IoT the big data concept has kept sidewise. Because big data and IoT has a very powerful association with each other. So many devices are used in the network to generate big data while pushing IoT concept towards the SIoT. The above mention study still has a deficiency of analyzing big data. So in IoT things require such technique through which big data can be analyzed and processed in an efficient manner. We believe that for the creek of big data processing to defining human dynamics there is an utter requirement of strong system architecture. IV.

THE SOCIAL STRUCTURE OF HUMAN BEHAVIOR

In this section, we provide the abstract level information about the social structure and their relationship with the human behavior. The role of social structure is referred to life-cycle of human being since it is the first form of parental relationship, i.e., parental object relationship [27]. Such relation is defined among analogous objects, which is constructed in the same period of time and by the same manufacturer. This, this relationship is easily entrenched during their new production. Also, it may not be changed over the period of time. Based on the social structure, human behavior is also coming from this relationship (referred to as statistical and physical complex systems). In addition, objects establish their relationship either by working together or working at the same place. It is similar to human’s way of life, e.g., human’s share personal experiences either secretly or in a public. Such types of relations are resolute whenever objects reside at the same place or they are interacted with each other using common medium, i.e., the Internet. These relationships are established as a location-based application or situation-based application [27]. Sometimes, changes are so frequent based on the duration of location or working, interaction frequency [28-29]. If we consider objects owned by a user, such mobile, laptop, etc. These devices are termed as ‘users-ownership’ as shown in Figure 1. Such type of ownership is generally for a longer period of time, and users are using these technologies for a longer period of time. These objects interact with another object with user choice, preferences, or sometimes due to the bound situation (e.g., emergency). These are devices are properly authorized by users and they are used to exchange data, e.g., contacts, messages, online purchasing, etc. The lashing theme is that devices with similar properties are used to share best practices in order to solve ‘friend’s problem’. Thus, having such knowledge of social relationships, the behavior can be derived from the interaction of these devices. The variations in behavior in terms of social and technological are the collections and values of billions or even trillions of individual user verdicts, benefits, activities, or primaries. These regularities in terms of variations and frequencies for human behavior are daily activities, such reading/sending emails, messages, facebook likes or sharing videos or other data,

sports, online shopping, and other work patterns are usually considered to be random. Also, the verdicts of these activities are difficult to judge/assess. Moreover, in 1837, Simeon Denis Poisson revealed that various activities follows up a distribution, i.e., denominated Poisson Distributed [30]. Thus, the discussion was made based on the traditional queuing model based on Poisson distribution, which offers new indications nearly the long tail distribution of inter events times/frequencies that unsurprisingly arise in human activities. Social Network

Service APIs

Social Graph

Tech. Network

RelationShip M ngt.

Participation model

Time based Location based

Situation based

other Network

Service Composition

Relation control

Owner influence

Human Behavior

Object influence

Other influence

Figure 1. Basic components of Social structure platform for human behavior

Usually, the change from burst or intensive activities to non-activities is a common nature of human being, who are tend to follow-up their routine fashion of diet, sports, online shopping, and social network in a very rigorous way up to a point where its priority is reduced or even forgotten. So far, the analysis of human behavior is examined by using weblogs, smartphones, shopping center queues, and sometimes interaction with physical stimuli i.e., earthquake, flood, and other disasters. Apparently, smart cities and the CPS opening a new way to a data source that enables new peers of consequences in enumerating and understanding of human behavior. Human behavior follows to comprehend through big data and evidence of non-Poisson distribution activity in an individual behavior. Nowadays, society is coming up with new ways of interactions between people via smart devices, social network, that lead towards an informative society, where information is just one click away from the user. Specifically, in this work, we analyze the potential of big data for smart cities, social network, and technical network through our proposed system architecture, and defining a new role by a novel terminology ‘socio-cyber network’. Secondly, analyzing big data that gives us a brief knowledge of the human behavior. And finally, influence on behavior via continuous feedback by means knowledge extraction and experiences. V. PROPOSED MULTILEVEL SYSTEM DESIGN The multilevel active storage and processing aim to extend the existing storage and processing systems. The system architecture is composed of four layers. Each layer is supported

by different functionalities enables read and write operations high effectively. In this section, we will first introduce the layered architecture that supports complete system design for high-performance computing. Afterward, we will present the design and working of the designed system. A.

IV-Tier Layered Architecture

Based on the needs of analyzing big data, we propose an IV-Tier architecture model. The designed model assists different objects to interact with each other using the shared medium. The proposed architectural model integrate various data generated by difference application, under the same domain, i.e., social internet of things, which supports the research community to provide the generalized framework and architecture that can help the domestic users in the case of security, healthcare, elderly age people and kids, and transportation system, machine-to-machine network, wireless sensor network, and vehicular network, etc. As Figure 2 shows that the proposed IV-Tier architectural model consists of four layers. Tier I: Data Generation handles data generation through various objects and then collecting and aggregating that data. Since a different number of objects are involved in generating the data. Therefore, an enormous number of heterogeneous data is produced with various format, a different point of origin, and periodicity. Moreover, various data have security, privacy, and quality requirements. Also, in sensor’s data, the Metadata is always greater than the actual measure. Therefore early registration and filtration technique are applied at this layer, which filters the unnecessary Metadata, as well as redundant data, is also discarded. Tier-II: This layer provides end-to-end connectivity to various devices. Moreover, data is aggregated at this point generated from various devices and arrange them in the proper format. Tier-III: Data storage and Processing Layer is the primary layer of the whole system architecture, which handles the processing of data. Since we need a real-time stream of the data and offline data analysis. Therefore, we need a third party real-time tool to combine with the processing server to provide the real-time implementation. To provide real-time implementations, Strom, Spark, VoltDb, and Hupa can be used. For instance, to be very specific in the case of data analysis, the implementation part could be achieved by using MapReduce. At this layer, the same structure of MapReduce and HDFS is used. With this system, we can also use HIVE, HBASE, and SQL supposed for managing Database (in-memory or Offline) to store historical information. Tier-IV: Service layer is the lowermost layer responsible for incorporating the third party interfaces to objects and human. This layer can be used autonomously as a single site, merged with other locations, or deployed in cloud interface. There are different other features as well. For instance, the unique global ID management is the key element in the application layer that handles identifying the object throughout the universe. Vendor control is another feature deals with the definition of the activities duly performed by different objects. The proposed architectural layers involve different objects that need intelligent power to interact with a human. For this reason, a

sm mart algorithm m is required aat the applicattion level thatt could effficiently and effectively innteract with tthe human. V Various tassks could be performed bby this featurees, such as request r

generator, sessionn initiating, setting up com mmunicating rrules, inteeract with heeterogeneous objects andd terminating the sesssion.

Figure 2. Foour-tier communnication model

Figure 3. Soocio-cyber systeem architecture

B.

Socio-Cyber Network

The idea of exploiting social network and technical network, termed as ‘socio-cyber network’, based on the Internet that allows objects to establish relation based on the extracting knowledge from the data. After extracting knowledge, the proposed system architecture generates the number of trails by looking into user behavior. Moreover, trust index is generated by finalizing specific trails of the experience, which helps in connecting graph of varies vertices of the experience as shown in Figure 3. In order to understand the actual working of the proposed system architecture, it is further divided into following sub-sections. 1. Data Collection Layer This layer is responsible for data generation. Devices involved in this layer uses communication medium, such as the Internet, ZigBee, Wi-Fi, etc. to communicate with each other. The communication medium handles collecting data from all the objects (i.e., social network, technical network, and another network) and then relay it towards the communication layer as shown in Figure 3. Initially, metadata is collected whose nature heterogeneous. Moreover, this layer is also capable of finding the redundant data. For this technique, some related techniques are used to find the redundant data [1] [31]. Afterward, those metadata, as well as redundant, are discarded. In the literature, various schemes tackle the entire amount of data, which uses additional system requirements. Therefore, the proposed system does not encounter processing of raw data, redundant data, or metadata. After elimination of redundant data, the useful amount of data is classified by the identifier and message type. Once the data is classified, the data is converted to machine readable form, which provides an easy solution for the processing to understand it very well and process accordingly. 2. Communication Layer This layer is responsible for transmission of data from source to the design system for analysis purpose. It used high-speed Internet, GPRS, 3G, 4G/LTE, or WIMAX as a source of the medium provider. In addition, it uses Wi-Fi or Bluetooth communication technology to transfer data from source to designed server if the devices and system are kept near to each other. All the communication with the various units of the analysis system is done by Ethernet. In this layer, we exploit the nature of graph that generates or updates each time when new data is added to the system. Initially, it creates a new graph but at later stages, when it meets any incoming data, it just updates the graph by either adding a new node, new edge or updating the weights on the edge. It uses an efficient searching mechanism, which uses indexing to search particular edge to be updated when required. Graph building layer also increases the efficiency of the system by making the graph be processed on multiple parallel servers simultaneously while dividing the graph into various independent mutually exclusive parts/subgraphs. The exploitation of graphs in the system assists the HPC to process the data efficiently. It is important to note that unlike the previous system for HPC, our designed system is not only deals with the processing efficiency. However, its main task is to analyze the huge amount of data in limited resources of the Hadoop server.

3. Processing Layer This layer is responsible for processing sub-graphs which were sent to the processing server by the intermediate layers. From the literature, it is recognized that the traditional approaches do not efficiently analyze Big Data. Therefore, a new system with novel algorithms is required that can efficiently analyze human behavior using Big Data. Therefore, keeping in view the aforementioned requirements, the proposed system architecture is powered by the processing layer, which acts as a core component for the designing system. The processing layer initially performs load balancing algorithm. Load balancing is used for distributing load to each server in equal size. This equal size distribution enhances the system efficiency and all the server will process the equal amount of output, and generate an output at the same time. After load balancing, the data chunks (referred to as sub-graphs) are sent to the raw data storage device. Raw data storage device is used for storing sub-graphs in a sequential form. This unit helps the system if the data has some missing values, it can re-accessed from the previous layers. After storing the data, if there is metadata (after arranging data in graphs form) than it is the time to discard those data. Moreover, the semantic engine will check whether the graphs that are ready for processing is at low scale, medium scale, or large scale. The semantic engine pass-on their instruction to the Hadoop system to react accordingly (i.e., assign memory and processing power). a. Knowledge Extraction: Knowledge extract is a process of capturing useful information, developing, sharing effective information, and recapitulate all activities with a motive of using knowledge in an efficient manner so that users get their required objectives. Generally, Knowledge Pyramid (DIKW Pyramid) presented in [32], is used for illustration functional relationship among data (D), information (I), knowledge (K), and wisdom (W). Therefore, based on the aforementioned explanation of DIKW, we consider the data and put together all the data in a context. When certain information becomes actionable, it is then transformed into knowledge. Afterward, when a certain information is consolidated, it transforms into the experience. Initially, the data is considered to be a subgraph. For instance, these sub-graphs include the data of temperature, traffic data, or other technical data. When a certain action is performed by sub-graphs, which become more sophisticated data. So, in other words, these sub-graphs include directions, which specifies the relationship between two entities, i.e., online data, temperature data, and other data. It can be stated in another word, e.g., we have a knowledge base. The scenario of the action (let say emergency alarm, online shopping, etc.) specifies how our system reacts in the modifying environment. Moreover, how a solution to the action is performed is achieved by analyzing sub-graphs since sub-graphs include insight information. Afterward, this level provides the advantage of learning mechanism using previous experience. After specifying knowledge from the data, the useful information is referred follow the friendship selection, which specifies the notion of trust index using graphs. b. Trust index based on friendship selection: At this section, the exploitation of geo-location information is used to ensure the friendship selection among various data attributes based

on the locatiion of the obj bjects. For insstance, if maxximum users reside in one housee, so it is likkely to happeen that majority of the users has probablyy similar chhoices. Moreover, thhey need similar environmeent behavior, which requires balaanced temperaature or moistture inside a house. Apparently, iif users are farr away from eaach other, thenn there is very less cchance to be ccorrelated witth each other due to high-securityy concerns. Inn this case, the trust inddex is drastically deecreased. For trust index, we are consiidering distance as a threshold. F For instance, if one user is i two kilometers aw way from anotther user, and there is no freequent calls, messagges, or any othher contact. Thhen, we can saay that both users doo not have any relation. Andd hence, there should s be no trust vaalue. Moreoveer, trust index can be calculaated as how many tiime users sharre informationn with anotheer user using Internet services.

Figure 4. Graph developpment for smart transportation

Based on thee aforementionned informatioon, the use off graph algorithms w with interchaanging weighhts is propossed to analyze humaan behavior baased on frienddship selectionn. Here we provide a description off few of the usse cases of thee smart wever, transportationn decisions ussing graph alggorithms. How the practice iss not limited oonly to these use cases; we ccan use graph technoologies and allgorithms to make lots off other transportationn-related com mplex decisioons. We aree also describing hoow the graphs are generated from vehiculaar data and few otheer smart trannsportation decisions that ccan be made using graphs, g such as, finding quuickest and shhortest path towards destination, ffinding the coongested or bllocked road, findinng the quickkest route too more thann one destinations, how a proposeed system willl guide user baased on the experiencce and knowleddge of the prevvious system etc., e by using real-tim me vehicular ddata.

edges, which is i denoted byy RG = ( , E , ) with tthree tyypes of weiights. Each intersection of the roadds is reepresented by a vertex of thhe road graph R RG, denoted bby , w where i repreesents the inntersection nuumber. The road bbetween two inntersections iss represented by an edge of the ggraph between two vertices, denoted by (E E , ), wheree E is thhe edge (road) from interrsection too ),. Each edge (E E , ),) havee three type oof weights, whhich representss the current traffic scenarios inccluding 1) the distance betw ween wo intersectioon and , ddenoted by DIST D , 2)) the tw aaverage speed of all vehicles from inteersection andd , ddenoted by AV VG_SP , 33) the numberr of vehicles ggoing fr from intersectiion to , ddenoted by NO_VEH N , . A saample road graph of a smaall part of thee city is show wn in F Figure 4. Thhe undirected edges show w the road from inntersection tto and from to . T The graph proocessing is doone by dividinng the graph into m mutually excluusive N subgrraphs, i.e., G11, G2, G3 …, Gn suuch that G1 ∩ G2 ∩ G3 ∩ … ∩ Gn = Φ, aas shown in Fiigure 55. It is better ooption to diviide the graph based on thee city environment, ii.e., dividing the graph depending uponn the bbridges in the ccity. Each of tthe Subgraph Gi is processeed by thhe separate noode and then at result agggregation level the reesults are agggregated. Lateer, the follow wing decisionss are m made using varrious graph allgorithms by pparallel processsing oof the subgraphhs. Now w the data is reeady for processing in the deesigned server.. The desiigned system iis powered paarallel algorithhms, which equually andd parallel proceess the data. O Once the data is processed,, it is requuired to the stoore the resultss in a local dissk, where theyy can be used u for futurre usage. Thuus, it is reallyy hard to storee the resuults or perform m a traditional MapReduce M fuunction. In ordder to cope with such siituation, we caame up with thhe extension oof the systtem that providdes enough loccal memory too store the dataa and resuult. For thiss purpose. T The incorporration knowlledge extrraction, reliabbility index, and a trust indeex. In some ccases usinng traditional aapproaches, soometimes we ddo not get whaat we actuually want, or the results arre corrupted. H Hence, we neeed to proccess the wholee data again. Thus, T reducing the computatiional andd processing efficiency off the system m. Therefore, our desiigned system sstores the resuults, analyze thhe results and then it is displayed to tthe users. VI. IMPLEEMENTATION RESULTS AND D ANALYSIS

Figure 5. Divisions of graphhs

Graph Buildiing: The traffi fic data is reprresented by diirected and weightedd road graph RG G including veertices and weeighted

Thee proposed sysstem is implem mented using S Spark and GraaphX withh Hadoop sinngle node seetup on UBU UNTU 14.04 LTS coreeTMi5 machinne with 3.2 GH Hz processor aand 4 GB mem mory. For real-time traaffic, we gennerated Pcap packets from m the dataasets by usinng Wireshark libraries andd retransmit tthem tow wards the system. Hadoop-pcapp-lib, developed Haddoop-pcap-serde, and Hadooop Pcap Inputt libraries are used for network paackets processing and geenerating Haddoop Reaadable form (sequence file)) at collectionn and aggregaation unitt so that it ccan be processsed by Hadooop and GrapphX. GraaphX is used tto build and pprocess graphhs with the aim m of makking smart trannsportation deecisions. We have consideredd the masssive volume of data from m [33-39].Thee intensity off the

Innitially, the analysis is perfoormed on Aarhhus city traffic. The sppeed analysis with respecct to the inteensity of trafffic is peerformed, show wn in Figure 88. When the iintensity of traaffic is m more, i.e., moree vehicles on the road betw ween two poinnts, the avverage speed of the vehiclles is greaterr. The fall in some veehicles on the road results inn a rise in thee average speeed. We caan easily notiice a higher nnumber of caars, i.e., 25-330, the avverage speed iss very low at vvarious times of o the day, shoown as a red color graaph. Whereas, at a lower intensity, i.e., 0-10 shhown as a bluee line, the averrage speed of the vehicles iss quite higher. There aare also some abnormalitiess exist with a lower nuumber of vehiccles the averagge speed is alsso lower. This might bee because of the constructtion of the rooads or somee other incidents. Norm mally, the distaance is conserrved to measuure the tim me to reach thhe destinationn. However, we w observed thhat the nuumber of vehiccles and the avverage speed aalso affects thhe time to reach the desstination. Figurre 9 shows thee blockage of one of the roads in Aaarhus city. Baased on the prroposed schem me, the avverage speed oof the vehicles is too low eveen when the nnumber off vehicles is also a low. W We can see thaat most of thee road blockage is donne at morningg times on diffferent days. T This is beecause of the rroad constructiion and workiing at morningg time. Thhe time to reaach the other ppoint is shownn in Figure 100, with respect to the inntensity of the ttraffic. We cann easily perceiive the r results inn more increase in the nnumber of vehhicles on the road me to reach tto the other ppoint. More vehicles on thee road tim reduce average speed of thee vehicles, whhich results inn more me to reach thhe destination.. As a result of o these phenoomena, tim

Figuure 11 shows the percentagge of humidityy inside the hoome. Gennerally, humiddity plays an im mportant role oon user behaviior in casee if the user is doing physicaal exercise or any other actiivity. Morreover, if therre is an increase in the hum midity, the usagge of elecctricity also inncreases. For this case, the proposed schheme exploits the phennomena of leaarning mechannism. Humidiity is meaasured by sennsors, and thhis data is trransferred to our propposed schemee for experienccing the levell of humidity. Our propposed schemee considers sevveral reading, and thus creaating one threshold durring the montth of Decembeer 2016. Baseed on wledge, the prroposed schem me will predicct for the previous know J 20166. Thus, the user will rreact the month of January wn in accoordingly if huumidity is incrreased or decrreased as show Figuure 11. Similarrly, the same ttechnique is foollowed for ouutside tem mperature as shhown in Figuree 12.

Figuree 6. Number of vehicles in giveen time 250

Le ess vhicals (1‐15)

More Vhicals (25‐35)

200

150

100

50

0

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101

Fiigure 7 shows the estimated time to reach home. For insstance, the user often travels from m home to ooffice. Initiallyy, the m experiencees the route and a gathers aall the prroposed system traaffic analysis based on the graphs. Afterrward, the prooposed syystem analyzees traffic data. Once dataa is processeed and annalyzed, it givves suggestionns to the user about the estiimated tim me to reach their home. Hennce, it dependss on user whetther he foollow such insttructions or noor. Thus, by dooing so, the user will haave accurate information about the tim mings to reaach its deestination. Mooreover, users at home cann also get the actual tim me of the userr to be arrivedd at home. It iss quite useful dduring the emergency scenario, e.g., if the user iss late then hiss usual mings, then it is possible thaat some emerggency has occuurred. tim

we take real-timee traffic inform mation to calcculate the shoortest andd quickest pathh between souurce and destiination rather than onlyy the distance information.

ESTIMATED TIME TO REACH (MIN) ( )

traaffic varies froom time to tim me on the samee road. The inttensity annalysis at the various v time oof the day helpps the authoriities to m manage and maake a proper plan for the trafffic on that parrticular me. Figure 6 sshows the inteensity of the ttraffic in one of the tim rooads of Aarhus city. We cann see at earlyy morning 7:00-9:00 annd noon time 11:25-12:30; the traffic is higher on thee road. Thhis might be bbecause of the office and schhool start time and at nooon the kid's scchool end timee. Therefore, tthe proposed ssystem annnounced the authorities w when the inteensity of the traffic increases on a pparticular road at any time off the day. Morreover, the system also has the capabiility to identifyy the blockagee of the rooad based on thhe current trafffic informatioon. The blockkage of the road can bee identified byy the number of vehicles annd the avverage speed. When the nuumber of vehicles is more on the rooad, and the avverage speed iss too low, this shows the bloockage off the road. Theerefore, based on such expeerience, the user will tryy to find a tim me at which theere is very lesss traffic.

TIME

Figuure 7. Estimatedd time to reach home h

Figgure 8. Averagee speed of a vehhicle

vehiclee Count

25 20 15 10 5 0

2014‐10‐01T05:40:00 2014‐10‐01T06:00:00 2014‐10‐02T05:55:00 2014‐10‐06T05:55:00 2014‐10‐08T05:40:00 2014‐10‐22T05:45:00 2014‐10‐23T05:45:00 2014‐10‐23T06:05:00 2014‐10‐27T06:50:00 2014‐10‐27T07:10:00 2014‐10‐28T06:40:00 2014 10 28T07:10:00 2014‐10‐28T07:10:00 2014‐10‐29T06:35:00 2014‐10‐30T07:00:00 2014‐11‐03T06:40:00 2014‐11‐04T06:45:00 2014‐11‐05T06:40:00 2014‐11‐05T07:00:00 2014‐11‐06T07:00:00 2014‐11‐10T06:40:00 2014‐11‐11T07:00:00 2014‐11‐12T06:50:00 2014‐11‐13T06:30:00 2014‐11‐13T07:00:00

Average Speed (km/h)

Average Speeed

20 18 16 14 12 10 8 6 4 2 0

Date aand Time

Fiigure 9. Averaage speed durinng various datte and time

The effect of processing tim me with respeect to increasinng in the graph is also examined whhile evaluatingg the efficienccy of the system. We tested the syystem by increasing numbeer of m zero to one hundred thoussand, noddes and numbeer of edges from as shown s in Figuure 14. The maassive increasee in the number of edges and nodes rresults in a grradual increasee in the processsing Moreover, eveen for one hunndred timee while buildinng the graph. M thouusand nodes and a edges, thee processing tiime is quite loower, i.e., less than onee thousand miilliseconds. Thherefore, baseed on the efficiency resuults, we can saay that the sysstem performs well X on andd in a real-timee if it is develooped using Spaark and GraphX m. Haddoop ecosystem

Figuure 10. Number of vehicles in ggiven time

120 100

Humidity (%)

Since the mainn contributionn of the work iss the processinng of largge graphs to acchieve smart trransportation, thus, the systeem is evalluated with reespect to efficiiency in termss of throughpuut (in meggabytes/sec M Mbps) and the rresponse timee (in milliseconnds). While analyzing the efficiencyy results in terrms of throughhput, we increase thee dataset sizze and perceeives the syystem throoughput effectts. We noticeed that with thhe increase inn the dataaset the system m throughput is also increaased, as show wn in Figuure 13. To suum up, we cann say the throoughput is directly propportional to thhe data rate. This is becauuse of the parrallel proccessing of the large graphs oon Hadoop ecoosystem. Whenn the dataaset is larger, the Hadoop system partittions the data into chunnks and proccess them in pparallel. Wee can examinee the throoughput at hhigher (larger) dataset i..e. 5345MB, the throoughput for thhis dataset is quite q better thhan other systtems. Thiss is the majoor achievemennt of the sysstem that withh an incrrease in data size the througghput is also inncreased. How wever for tthe smaller daataset. Less thaan 100 MB, thhe use of Hadoop is not efficient.

80 60 40 20 0 0

5

10

15 5

20

25

30

35

Date (JJanuary 2016)

High

Avg

Low

Figure 11. Hum midity inside hoome 20

Figure 13. Thrroughput of the system depending on data sizee

Temperature (0C)

15 10 5 0 0

5

10

15 5

20

25

30

35

‐5 ‐10 ‐15

Date (January 2016)

High

Avg

Low

Figure 12. Ouutdoor temperatture

Figgure 14. Graph ggeneration time depending on tthe number of edges

VII. CONCLUSION In this paper, we proposed a novel notion ‘socio-cyber network’ for defining human behavior using big data analytics. Generally, big data is generating by smart devices that tend to communication via the Internet. Generally, human behavior is defined by exploiting social network, such as smartphone, and social network. However, it does not provide in-depth knowledge of information resides inside the data. Therefore, based on aforementioned features, we proposed a scheme based on the experiencing the data based on learning mechanism. Such experienced is gained by finding the friendship level among devices. Friendship level is further enhanced by exploiting trust index, which eventually lead toward the notion of graph theory concept. In our proposed scheme, graph concept is introduced that helps the devices to find a better match and can predict the future. After processing the data using GraphX, results are finally stored in the results storage device in the system architecture. The user can utilize these results and can predict their future behavior, which helps in traffic management system, indoor and outdoor temperature. The performance of the system architecture is tested on Hadoop using UBUNTU 14.04 LTS coreTMi5 machine with 3.2 GHz processor and 4 GB memory. The final evaluations show that the performance of the proposed network architecture fulfills the required desires of the users connected to it, whether the input data is a real-time as well as offline while taking actions at the real time.

[5] [6] [7]

[8]

[9] [10]

[11] [12]

ACKNOWLEDGEMENT

[13]

This work was supported by the faculty research fund of Sejong University during 2017-2018. This research was also supported by Next-Generation Information Computing Development Program through the National Research Foundation of Korea(NRF) funded by the Korea government(MSIT) (2017M3C4A7066010)

[14]

[15] REFERENCES [1]

[2]

[3]

[4]

Awais Ahmad, Anand Paul, and M. Mazhar Rathore. "An efficient divide-and-conquer approach for big data analytics in machine-to-machine communication." Neurocomputing 174 (2016): 439-453. Awais Ahmad, Anand Paul, M. Mazhar Rathore, and Hangbae Chang. "Smart cyber society: Integration of capillary devices with high usability based on Cyber– Physical System." Future Generation Computer Systems56 (2016): 493-503. Guinard, Dominique, et al. Towards physical mashups in the web of things. Networked Sensing Systems (INSS), 2009 Sixth International Conference on. IEEE, 2009. Jara, Antonio J., Yann Bocchi, and Dominique Genoud. "Social Internet of Things: The potential of the Internet of Things for defining human behaviours." In Intelligent Networking and

[16]

[17]

[18]

Collaborative Systems (INCoS), 2014 International Conference on, pp. 581-585. IEEE, 2014. Sanchez Pineda, A., Maranon-Abreu, R. Like Art: Integrating Internet of Things and Social Networks, http://like-art.com/, 2014. D. Guinard, M. Fischer, and V. Trifa, “Sharing using social networks in a composable web of things,” in Proc. IEEE PERCOM, 2010. H. Ning and Z. Wang;, “Future Internet of things architecture: like mankind neural system or social organization framework?” IEEE Commun. Lett., vol. 15, no. 4, pp. 461–463, 2011. M. Kranz, L. Roalter, and F. Michahelles, “Things that Twitter: social networks and the Internet of things,” in What can the Internet of Things do for the Citizen (CIoT) Workshop at Pervasive, May 2010. Dixit, Sudhir, and Ramjee Prasad, eds. Technologies for home networking. John Wiley & Sons, 2007. Nitti, Michele, Luigi Atzori, and Irena Pletikosa Cvijikj. "Friendship selection in the social internet of things: challenges and possible strategies." IEEE Internet of Things Journal 2, no. 3 (2015): 240-247. T. Sun, J. Deng, and K. Deng, “Scale-free network model with evolving local-world,” in Proc. 4th Int. Nat. Comput. Conf., 2008, vol. 1, pp. 237–240. K.-K. Yap, V. Srinivasan, and M. Motani, “Max: Human-centric search of the physical world.” in Proc 3rd Int. Conf. Embedd. Netw. Sensor Syst. (Sensys’05), 2005, pp. 166–179. Dean, Jeffrey, and Sanjay Ghemawat. "MapReduce: simplified data processing on large clusters." Communications of the ACM 51, no. 1 (2008): 107-113. Ghemawat, Sanjay, Howard Gobioff, and Shun-Tak Leung. "The Google file system." In ACM SIGOPS operating systems review, vol. 37, no. 5, pp. 29-43. ACM, 2003. Ranger, Colby, Ramanan Raghuraman, Arun Penmetsa, Gary Bradski, and Christos Kozyrakis. "Evaluating mapreduce for multi-core and multiprocessor systems." In 2007 IEEE 13th International Symposium on High Performance Computer Architecture, pp. 13-24. Ieee, 2007. Yoo, Richard M., Anthony Romano, and Christos Kozyrakis. "Phoenix rebirth: Scalable MapReduce on a large-scale shared-memory system." InWorkload Characterization, 2009. IISWC 2009. IEEE International Symposium on, pp. 198-207. IEEE, 2009. Rafique, M. Mustafa, Benjamin Rose, Ali R. Butt, and Dimitrios S. Nikolopoulos. "Supporting MapReduce on large-scale asymmetric multi-core clusters." ACM SIGOPS Operating Systems Review 43, no. 2 (2009): 25-34. S. Uppoor, O. Trullols-Cruces, M. Fiore, J.M. Barcelo-Ordinas, Generation and Analysis of a Large-scale Urban Vehicular Mobility Dataset, IEEE Transactions on Mobile Computing, Vol.13, No.5, May 2014

[19]

[20]

[21]

[22]

[23]

[24] [25] [26]

[27]

[28]

[29] [30] [31]

Ning, Huansheng, and Ziou Wang. "Future Internet of things architecture: like mankind neural system or social organization framework?." Communications Letters, IEEE 15, no. 4 (2011): 461-463. M. Mazhar. Rathore, A. Paul, A. Ahmad, B.W. Chen, B. Huang, Wen Ji "Real-Time Big Data Analytical Architecture for Remote Sensing Application", IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Volume: PP, Issue: 99, Page(s): 1 – 12, Digital Object Identifier 10.1109/JSTARS.2015.2424683, 1939-1404 © 2015 IEEE. Chung, T.Y., Mashal, I., Alsaryrah, O., Chang, C.H., Hsu, T.H., Li, P.S. and Kuo, W.H., 2014, September. MUL-SWoT: A Social Web of Things Platform for Internet of Things Application Development. In Internet of Things (iThings), 2014 IEEE International Conference on, and Green Computing and Communications (GreenCom), IEEE and Cyber, Physical and Social Computing (CPSCom), IEEE (pp. 296-299). IEEE. C. Zhang, C. Cheng, and Y. Ji, “Architecture design for social webof things”, in Proceedings of the 1st International workshop on Contect Discovery and Data Mining, 2012, p. 3. D. Duinard, M. Fischer, and V. Trifa, “Sharing using social networks in a composable web of things”, in Pervasive Computing and Communications Workshop (PERCOM Workshops), 2010 8th International Conference on, 2010, pp. 702-707 T. Schmid and M. B, Srivastava, “Exploiting social networks for sensor data sharding with SenseShare”, 2007. Labrinidis A, Jagadish HV “Challenges and opportunities with big data”, Proc VLDB Endowment 5(12):2032–2033, 2012. Le, Xuan Hung, Sungyoung Lee, Phan Tran Ho True, Asad M. Khattak, M. Han, D. V. Hung, Mohammad M. Hassan et al. "Secured WSN-integrated cloud computing for u-life care." In Proceedings of the 7th IEEE conference on Consumer communications and networking conference, pp. 702-703. IEEE Press, 2010 Atzori, Luigi, Antonio Iera, and Giacomo Morabito. "Siot: Giving a social structure to the internet of things." IEEE communications letters 15, no. 11 (2011): 1193-1195. L. E. Holmquist, F. Mattern, B. Schiele, P. Alahuhta, M. Beigl, and H. Gellersen, “Smart-its friends: a technique for users to easily establish connections between smart artefacts,” ACM Ubicomp’01. A. P. Fiske, “The four elementary forms of sociality: framework for a unified theory of social relations,” Psychological Review, vol. 99, 1992. Haight, F. A., Handbook of the Poisson distribution. 1967. Ahmad, Awais, Anand Paul, Mazhar Rathore, and Hangbae Chang. "An efficient multidimensional big data fusion approach in machine-to-machine

[32] [33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

communication." ACM Transactions on Embedded Computing Systems (TECS) 15, no. 2 (2016): 39. J. Rowley, The wisdom hierarchy: representations of the DIKW hierarchy, Journal of Information Science 33(2) (2007) 163-180. Stefan Bischof, Athanasios Karapantelakis, Cosmin-Septimiu Nechifor, Amit Sheth, Alessandra Mileo and Payam Barnaghi. 2014. Semantic Modeling of Smart City Data. Position Paper in W3C Workshop on the Web of Things: Enablers and services for an open Web of Devices, 25-26 June 2014, Berlin, Germany. R. Tönjes, P. Barnaghi, M. Ali, A. Mileo, M. Hauswirth, F. Ganz, S. Ganea, B. Kjærgaard, D. Kuemper, S. Nechifor, D. Puiu, A. Sheth, V. Tsiatsis, L. Vestergaard. 2014. Real Time IoT Stream Processing and Large-scale Data Analytics for Smart City Applications. poster session, European Conference on Networks and Communications 2014. Sefki Kolozali, Maria Bermudez-Edo, Daniel Puschmann, Frieder Ganz, Payam Barnaghi. 2014. A Knowledge-based Approach for Real-Time IoT Data Stream Annotation and Processing. In Proc. of the 2014 IEEE International Conference on Internet of Things (iThings 2014), Taipei, Taiwan, September 2014. Gramaglia, M., Trullols-Cruces, O., Naboulsi, D., Fiore, M. and Calderon, M. 2014. Vehicular networks on two Madrid highways. In 2014 Eleventh Annual IEEE International Conference on Sensing, Communication, and Networking (SECON) (pp. 423-431), 3 July, 2014, Singapore. S. Uppoor, O. Trullols-Cruces, M. Fiore, J.M. Barcelo-Ordinas. 2014. Generation and Analysis of a Large-scale Urban Vehicular Mobility Dataset. IEEE Transactions on Mobile Computing, Vol.13, No.5, May 2014. D. Naboulsi, M. Fiore. 2013. On the Instantaneous Topology of a Large-scale Urban Vehicular Network: the Cologne case. ACM MobiHoc 2013, Bangalore, India, July 2013 S. Uppoor, O. Trullols-Cruces, M. Fiore, J.M. Barcelo-Ordinas. 2014. Generation and Analysis of a Large-scale Urban Vehicular Mobility Dataset. IEEE Transactions on Mobile Computing, Vol.13, No.5, May 2014. Awais Ahmad, M. Khan, A. Paul, S. Din, M. M Rathore, G.Jeon, G. S. Choi “Towards Modeling and Optimization of Features Selection in Big Data based Social Internet of Things.” Future Generation Computer System 2017. A. Paul, Awais Ahmad, M. M. Rathore, and S. Jabbar “SmartBuddy: Defining Human behaviors Using Big Data Analytics in Social Internet of Things.” IEEE Wireless Communications Magazine, October 2016. Sadia Din; Awais Ahmad; Anand Paul; Muhammad Mazhar Ullah Rathore; Gwanggil Jeon, “A Cluster-Based Data Fusion Technique to Analyze Big Data in Wireless Multi-Sensor System.” IEEE Access Year: 2017, Volume: 5 Pages: 5069 – 5083.

[43]

[44]

Awais Ahmad, A. Paul, S. Din, M. M Rathore, G. S. Choi, G. Jeon “Multilevel Data Processing using Parallel Algorithms for Analyzing Big Data in High Performance Computing.” International Journal on Parallel Programming, (2017) Springer. pp 1–20, DOI: 10.1007/s10766-017-0498-x. Sadia Din, Anand Paul, Awais Ahmad, Jeong Hong Kim “Energy efficient topology management scheme based on clustering technique for

software defined wireless sensor network”, Peer-to-Peer Network and Applications. (2017). https://doi.org/10.1007/s12083-017-0607-z

Biograp phy: Dr. Awais Ah D hmad receiveed his Ph.D. in Computer Science aand Engineerring from K Kyungpook Naational Univer ersity, Daegu, Korea. He iss currently w working as an Assistant Prrofessor in th he Departmentt of Informatiion and Comm munication E Engineering, Yeungnam Y U University. In 2014, he wass also a visitiing researcher in INTEL-N NTU, Nationaal Taiwan U University, Taiiwan, where hhe was workin ng on Wukong g Project (Smaart Home). Siince 2013, hee has publish hed more thhan 65 Intern national Journ nals/Conferennces/Book Ch hapters in vaarious reputed d IEEE, Elsevvier, and Sprin nger Journals, whereas in leeading confereences, i.e., IE EEE Globecom m 2015, IEE EE Globecom 2016, IEEE LCN, 2016, and IEEE IC CC 2017, respectivvely. Dr. Awaais is also serrving as Guesst editor in various v Elsevier and Springger Journals, including FGCS, S Sustainble Cissties and Sociieties, and RT TIP. Moreoveer, he is an in nvited review wer in variouss journals, includingg IEEE Comm munication Leetters, IEEE T Transactions on o Wireless Communicatio C ons, IEEE Traansactions on Intelliigent Transpoortation System m, IEEE JSTA AR, Ad-Hoc Networks Elssevier, Compuuter Network,, Elsevier, and IEEE E Communicaations Magaziine. Dr. Ahmaad was the reccipient of threee prestigious awards: (1) IEEE I Best Researchh Paper Awardd: Internation nal Workshop on Ubiquitou us Sensor Sysstems (UWSSS 2015), in co onjunction with the Smart Worldd Congress (SWC 2015), B Beijing, Chinaa, August, (2) Research Aw ward from Preesident of Bahria U University Islaamabad, Pakisstan in 2011, (3) best Paper Nomination n Award in W WCECS 2011 at a UCLA, USA, andd (4) best Papper Award in 1st Symposiuum on CS&E, Moju Resortt, South Koreea, in 2013. His H current research interest includdes Big Data, Internet of Thhings, Social Internet of Th hings, and Hum man Behaviorr Analysis using Bigg Data. Dr. Aw wais has been n a key contribbutor in the saaid fields. He was also servving as a Lab Admin of CCMP L Labs from 2013 to 2017. Hee was also awaarded as Best Outgoing Ressearcher of CC CMP labs.

Muhammad Baabar receivess his Bachelo M ors in Compu uter Sciences with distincttion from U University of Peshawar, P Pakkistan in 200 08. He did his Masters of Sciences in Computer Sooftware Engiineering from m National University U Scciences and Technology (NUST), Isslamabad, Pak kistan in 20112. Currently y he is pursu uing his PhD D degree in Computer Sooftware Engiineering from m National University U Scciences and Technology (NUST), Isslamabad, Pak kistan. His reesearch area includes i but not limited tto Big Data Analytics, A Innternet of Thiings (IoT), S Smart City Design and Plaanning, and SSocial Web of o Things (S SWOT).

Saadia Din receeived her Bacchelors and Masters M in Co omputer Engiineering from m Comsats Innstitute of In nformation T Technology Abbottabad A and Abasyn University Islamabad, Paakistan in 2010 and 2015,, respectively. In 2015, she was a visititing research at CCMP Labs, Kyungpo ook National University Korea K where she was work rking on Big Data and Innternet of Thiings. She wass also workin ng as a Lecturrer in Abaysnn University Islamabad I unntil 2016. Currently, she iss doing here Master’s com mbined PhD pprogram in Ky yungpook N National Univeersity since 22016 March. Her area of research is 5G networkss and IoT Green commuunication, Big g Data analyttics, Wireless Sensor Netw work. She hass published feew highly enabled G reputed cconference succh as IEEE LC CN, ACM SA AC, ICC, Glob becom and some SCIE jourrnal at the beg ginning of her reseaarch career. In IEEE LCN 2017in Singapoore, she has ch hair couple off sessions. d is received his Ph.D. fro om University y of Manchestter, UK in 20 009. He is Shhehzad Khalid A Associate Profeessor at Deparrtment of Com mputer Engineeering, Bahriaa University, Islamabad I Paakistan

D Anand Pa Dr. aul, received the Ph.D. deg gree in Electrrical Engineerring from thee National Cheng Kung University, U Tain inan, Taiwan, in 2010. He is i currently w working as an Associate Prrofessor in th he School off Computer Science and Engineering,, Kyungpook National U University, Sou uth Korea. Hee is a delegatee representing South Koreaa for M2M focus group annd for MPEG G. His researchh interests incclude Algorith hm and Archiitecture Recon nfigurable Em mbedded Co omputing. Hee is IEEE Senior S memb ber and has guest edited d various innternational jo ournals and hhe is also part p of Editorrial Team forr Journal of Platform ysical Systemss. He serves ass a reviewer for fo various Technoloogy, ACM Appplied Computting review annd Cyber–Phy IEEE /IE ET/Springer annd Elsevier jo ournals. He is the track chaair for Smart human h compuuter interaction n in ACM SAC 20115, 2014. He was the recip pient of the O Outstanding In nternational Student S Schollarship award in 2004– 2010, thee Best Paper Award in Nattional Compuuter Symposiu um and in 200 09, and Internnational Confference on Softcompputing and Neetwork Securitty, India in 20015.

Dr. Mazzhar Ullah ratthore receiveed his Masterr degree in Computer C Co ommunicationn Security fro om the N attional University of Scieccne and Tech hnology, Islam mabad Pakista tan in 2012. Recently, hee completed his Ph.D. froom Kyungpook National University D Daegu. His reesearch in clludes Big Daata Analytics , Network Traffic T Analyssis andMonittioring, and Computer C annd Networ Seecurity.

Alavalapati Go A outham Reddyy is an Assisttant Professorr of the Deparrtment of Com mputer & Innformation Seecurity at the Sejong Univ versity, Repub blic of Korea. Before that, he was a V Visiting Reseaarch Scholar at the KIN NDI Center for Computiing Research at Qatar U University, Qaatar. He obtaiined his Ph.D D. in Computeer Science & Engineering g from the K Kyungpook Naational Univerrsity, Rep. Off Korea, in February 20177. His primary y research innterests revolv ve around Innformation Seecurity and IoT. I He holds ds several pu ublications inn cryptographiic authenticatiion protocols. He is a professional membber of the ACM M and the IE EEE.

Nasro Min-Alllah very succcessfully weaars many hatss: he is assoociated with CSAIL N C at M Massachusetts Institute of Technology as a Visiting g Scientist/Fac aculty; he is Associate Prrofessor (on leave) and thhe Computer Science Dep partment at C COMSATS In nstitute of Innformation Teechnology; he is the Director at Green Computing C andd Communicaation Lab. H received his He h Undergraaduate and Master M degreees in Electroonics and In nformation Teechnology in 1998 and 20001 respectivelly from Quaid di-Azam Univversity and Haamdaramd U University, Pak kistan. He obbtained a PhD D in Real-tim me & Embedd dded Systems from the G Graduate University of the C Chinese Acad demy of Scien nces (GUCASS), P.R Chinaa in 2008. He has bbeen enjoyingg a distinguish hed carrier booth in research h and academ mics. He is thee author of 35 5 research articles aand book chaapters publisheed in rankingg conferences, impact facto or journals annd book seriees such as Springer,, Elsevier, and John Wiley y & Sons etc.. His research h interests incclude softwaree pipelining, chromatic schedulinng, reliable coomputing and real-time systtems. He is allso the winnerr of followingg three most prestigious p awards: ii)-CIIT Goldeen Medallion for f Innovationn (CIMI-2009 9), ii)- Best Mobile Innovatiion in Pakistaan (BMIP2010), annd iii) Best Unniversity Teaccher Award, P Pakistan (2012 2).

Highlights    

Integrating CPS with Social Network Friendship selection based on neighboring devices Trust management based on geo-location information Analyzing Big Data using MapReduce and GraphX