Thispaper provides how Big Data Analytics, Cloud and Fog Computing to support inInternet of Things (IoT). The Big data analytics needed smart and efficientstorage. Fog computing extends the cloud computing paradigm to edge of thenetwork. If completely Big data analytics, Cloud and Fog computing are implemented in IoT than every aspect of human life is continuouslychanging. Fog and cloud computing provides everything as a services due tosocial media and smartphones have dominated our daily lives. The IoT plays an important term to describe theextension of connectivity and smart features in the year of 2020, to commonlyused appliances in our lives. Applicationsand services of IOT to get smart cities, smart homes, smart buildings etc..
Inthis paper we explore the relation between the IoT and emerging technologies likeBig Data Analytics, Cloud and Fog Computing. In addition to that we simulatethe protocol translator in application layer of IoT architecture. Keywords—Cloudcomputing; Big Data Analytics; Fog computing; Internet of Things. I.
Introduction The Internet of Things (IoT) provides latest development of RadioFrequency Identification (RFID), internet protocols, smart sensors,communication technologies and it improve quality of our lives1. Theapplications of IoT are transportation, health care, industrial automation, thepresent revolution of internet is mobile and machine-to-machine (M2M)technologies 1. The IoT permits physical objects to see, hear, think andperform jobs by having them talk together, the IoT transforms the objects tounderlying technologies are ubiquitous computing 1. The IoT providessignificant of human live and quality it deals with business applications andit grow the world economy1,2. Supposefor a example if smart city is made up of IoT then every home will enable their residents to automatically openthe door at the entrance when theirreaching home, prepare the coffee and breakfast automatically 1. To performthe climate control system and pervasive and ubiquitous nature it required tosupport emerging technologies 1. The emerging technologies which aresupporting to IoT are Big data Analytics, Cloud and Fog computing etc. The Big data can store massive data in theform of data, emails, texts, consumertransactions, social media interactions etc.
which is connected by a largenumber of devices and physical objects like humans, sensors, animals, plants,smart phones, personal computers etc. equipped with sensors to generate 1. The support of Big data analytics in IoT is achallenging task of organizations. In industries has required to process amassed volumes of data, any industryprocessed the amassed data with big data analytics three important factors i)Save the budget ii) Connect disparatesources of data within the business iii) Improvement of the revenue 1. In IoTthe use of big data analytics will create more accurate market strategies andenhanced Return on Investment (ROI) in the future sales2. Thecloud computing pay as you go model, it consist of millions of data centers andit composed of trillions of virtual machines (VMs).
The data centers areperforming higher utilization of VMs without degrading the performance. Theeffective tasks allocation strategy performing on VMs. In internet deployments,most notably the IoT, requires mobility support and geo-distribution andlocation awareness in addition to that low latency of plat form is fogcomputing 2.
There are seven characteristics of Fog computing there are i) Low latency and location awareness ii) Wide-spread geographicaldistribution iii) Mobility iv) Very large number of nodes v) Predominant roleof wireless access vi) Strong presence of streaming and real time applicationsvii) Heterogeneity 1. In Cloud and Fog computing provides developers andproviders to try everything together what customer wants. Cloud and Fog isessential to success of IOT, The following difference have made to define therole of IoT 1. Table 1: Cloud computingversus Fog computing in IoT Cloud Computing Fog Computing There are huge Data and applications are processed in a cloud, which is time consuming task for large volume of data in IoT 1. Rather than presenting and working from centralized cloud, fog operates on network edge. so it consumes less time in IoT 1 There is a Problem of bandwidth in IoT, as result of sending every bit of data over cloud channels is a tedious task1. Less demand for band width, as every bit of data’s where aggregated at certain access points instead of sending over cloud channels1. It was a slow response time and scalability problems as a result of depending servers that are located at remote places2.
By setting small servers called edge servers in visibility of users, it is possible for a fog computing platform to avoid response time and scalability issues1. The IoT interconnecting millions and billions or trillions ofconnections with sensors or large volume of data through the internet. In thisscenario the IoT layer architecture is drawn in the Figure 1, The IoT layeredarchitecture is a flexible and There are four IoT architectures had drawnFigure (a) baisc architecture modelconsisting of Application layer, Network layer and Perception layers. Figure(b), (c),(d) are illustrated the commonarchitecture of the five layer TCP/IP model, then it follow the briefdiscussion of some layersFigure 1: Thearchitecture of IoT 1 (a) Threelayered (b) Middle ware based layered (c) SOA based layered (d) Five layered(TCP/IP) Perception Layer: This is first layer of the three layeredarchitecture, the big data created by the IoT and initiated at this layer. Inthis layer digitizes the data and transfers to the Network layer through thesecure channels1. Object Layer: The objects layer is first layer of the SOA and Fivelayered which is also called perception layer or devices layer. The physicalsensor devices are connected to the objects layer. The object layer containssensors and actuators, it performing the various functionalities such as temperature humidity,weight, motions, vibration1.
Object AbstractionLayer: This is second layer of the SOA and Five layered, it transfers thedata to the objects layer and service management layer through the securechannels1.Service Management Layer:Service Management Layer is called as Middleware or pairing layer, this layerreceives the data and make the decisions and delivers the required servicesover the network 1.Application Layer:The application layer provides the environment of the services requested by thecustomer. The customer needs temperature, humidity and high quality of smartdevice1. Business Layer : Thebusiness layer is also called as management layer it manage the overall systemactivities and services, this layer builds the responsibility of the businessmodel such as graphs, statistical analysis etc. it receives the data andprocess the data based on the big data analytics and making decisions1.
Network Layer: In this layer transferthe data to IoT platform and performing the important role in IoT and wirelesssensor Networks (WSNs)1.The five layer model is most applicable model in IoT applications 1.There are six elements of combinationsare needed to build the functionalities of IoT identification, sensing, communication, computation, services andsemantics as shown in F igure 2.Figure 2: Theelements of IoT 1Identification is the first element it allocate the uniqueaddress of the objects for identifying unique objects. The addressing modeswhich are using identification isIPv4/IPv61.
Sensing is the secondelement combination of IoT, it gathering the data from related objects withinthe network and sending it back to a data base or data ware house or cloud 1.Communication element providescommunication protocols used for IoT are WiFi, Bluetooth, IEE 802.15.4, Z-wave 1.
Computation element processing themicrocontrollers, microprocessors, SOCs, FPGPAs and utilizing to provide theIoT hardware or software functionalities 1. Services Under the services of IoT can be categorized as four classes Identity-relatedServices, Information Aggregation Services, Collaborative- Aware Services andUbiquitous services are the main service 1. These services are useful tobuild smarthome, smart buiding, Intelligent Transportation System (ITS),Industrial automation, Smart health care, SmartGrid and Smart city 1. II. The Support Of Big Data Analytics, Fog AndCloud Computing Platformin IoT TheIoT is basically a complex network that seamlessly connects people and thingstogether through the Internet which is incorporated with Bigdata Analytics, Fogand Cloud computing. Theoretically, anything that can be connected through theM2M technologies (smart watches, cars, homes, thermostats, vending machines,servers) will be connected in the near future using sensors and RFIDtags 2. IoT will have the advantage of bringing us smart cities with smartcars, secure and efficient buildings, and smart traffic management systems 3.
It will achieve major efficiency in industry, healthcare and retail and willsave millions of dollars 3. IoT allows connected objects to continuously senddata over the Web and from anywhere 3. A. Bigdata Analytics platform The companies need to work with big data innovate their business, processing data, runsmarter applications and to deliver new values to their customers3. The datacollected through IoT becomes more useful if collected from different types ofdevices and then combined in a creative way3. This should be followed bybuilding connections and correlations between data units that lead tointelligent decision making processes3.There are various software development tools are used for Big data.
Hadoop and its incorporated technologies such as HDFS,MapReduce, Hive, Pig and others that support Big data paradigms that willembrace a large part of the IoT functionality1. There are threechallenges are involved in processing the data, first one is data collection itgathering data from different sources IOT connects all manner of end points,second one is data storage which is several possible directions in this regardsuch as storing the data from analytics in a relational database, in the cloud or in a NoSQL database MongoDBand CouchDB, third one is data analytics and business intelligence toolsempower decision markers as never before by extracting and presenting meaningfull information in real time 1. The relation between IoT, Big data and Big dataanalytics as shown Figure 3. The IoT connects end points to separation of dataand it extract value from the huge amount of collected big data and analytics. Thisrequires the development of applications that can analyze the data 4.
Figure 3: The role of Big data Analytics inIoT4The main advantage of Big Data to IoT is that predictive analytics isprovided over all the data, not only a small part of it. This allows to dig upfor patterns, correlations, and build insights from data stored in Big Datadatabases in a way never expected before. For IoT, the use of Big Data willcreate more accurate market strategies and enhanced Return on Investment (ROI)in the future sales. This is because data analysis will be implemented over thecomplete product lifecycle, and we will get feedback from devices as well asfrom customers.
B. Fog and Cloud computingplatform In Fog Computingcludlets are edge computing, fog is highlyvirtualized by supplementingthe cloud and providing intermediate layers of computation, networking, andstorage, fog nodes can optimize Internet of Everything (IoE) deploymentsgreatly enhancing latency, bandwidth, reliability, security, and overall IoEnetwork performance. Fog can be act as bridge between the smart devices. Fogand cloud perform the massive computational, storage capabilities. Fog andcloud perform the IoT Services to the end users. There are fog computing havingthe various features Location, Distribution, scalability, density of devices,Mobility support, Real-time, Standardization, On the fly analysis. Fog andcloud computing perform the big data analytics through IoT edge devices asshown in figure 4.
Figure4: The role of Big Data Analytics, Fog and Cloud in IoT services 1 In Figure 4 ,hundreds of cloud resources are data centers, datacenters interacted withthousands of fog gateways interacted with IoT services. IoT services deliversthe millions of services to end users. Big data analytics, Fog and Cloudcomputing services are provide everything as service through mobile networkoperators are at the cell towers.
The role of fog and cloud-basedarchitecture of IoT where application intelligence and storage are centralizedin server side data centers, satisfies the need of most of the IoTapplications. III. Analytics Of IoT And Interplay Between Big DataAnalytics, Fog And Cloud Computing The IoE billionsand trillion opportunity over past 10years. Eight hundred million devices will be connected to various networks in2020 as per the intelligence estimate statistical reports as shown in Figure 55. To bringing new technicalchallenges in big data analytics, Fogand cloud domains and specifically the rate may increase in the data processingthe number of devices which are connected . Big data contains hugeamount of data, The IoT is allowing us to generate more data with big dataanalytics, and it will generate the eye- popping numbers. The IoT which consists of all people and thingsconnected to the Internet, will generate 507.
5 zettabytes of data by 2019, accordingto Cisco one zettabyte is equal to onetrillion gigabyte 7,8. Figure 5: The deviceswhich are connected to IoT 5, 6 BI Intelligencereport believes that fog computing will be instrumental in analyzing all data,as it offers several advantages that a cloud computing model, itinclude quicker data analysis, reduced costs tied to data transmission,storage, and management, as well as enhanced network and applicationreliability 7. Fog computing is extending of cloudcomputing, it also involves delivering data, applications, photos, videos, andmore over the Internet to data centers 3.
IoT is a connection of devicesto the Internet, automobiles, kitchen appliances, and even heart monitorscould all be connected through the IoT 6. IoT in the future large numberof devices will join that list. The IoT generates massive amounts of data, andfog and cloud computing provides a pathway for that data to travel to itsdestination 7. Some of the most popularIoT cloud platforms on the market include Amazon Web Services, GE Predix,Google Cloud IoT, Microsoft Azure IoT Suite, IBM Watson, and Salesforce IoTCloud 7,8. Fog computing is aclever name it also known as edge computing, it provides a way to gather andprocess data at local computing devices instead of in the cloud or at a remotedata center 7. Under this model, sensors and other connected devices senddata to a nearby edge computing device 7.
This could be a gateway device, such as a switch or router, thatprocessors and analyzes this data 8. This is not just aboutaggregation or concatenation of sensed physical data like a typical gateway butreally about distributed intelligence, where effective real time anddeterministic processing is needed to implement a functionality 8. To movefrom cloud computing, or centralized computing, to edge computing according toa 2014 Cisco survey and 2017, the 37percent of IoT computing will be located at the edge of the network as shown inFigure 5 & 6 7,8.
Figure 6: Thegovernment and enterprises IOT devices connected to an edge solution 7, 8 IV. The Simulation Of Protocol Translator Provides Application Layer In gateway entity within the context of IoTnodes can be deployed by the thousands or even millions in support of a singleapplication. Thus, having self-management Fault, Configuration, Accounting,Performance and Security (FCAPS) capabilities is a must1,2.
In IoT process thecommunication over long distances between different systems, a range ofcommunication protocols are involved in such as Wi-Fi, Bluetooth, GPRS, 3G,LTE, ZigBee, Z-Wave, home automation communication protocol, Near FieldCommunication (NFC) which is an ensemble of protocols that allow electronicdevices to establish radio communication either by touching them together or bybringing them into proximity and many other forms of data connectivity 3. TheIoT devices can be classified as two major categories first one isresource-constrained and resource-rich devices. Most of IoT applications arelow-rate but the large number of IoT devices participating on a singleapplication needed gate way protocols.
We believe that there is a re-programmabilityof the IoT gateway through a rule-based language can put the gateway in aunique position to offer smart autonomic management, data aggregation or flowaggregation, and protocol adaptation services. There is a huge IoT load among the gateways it become multiple gatewaysit required unique solution. Itneeds an efficient solution for protocol conversion, it required aprotocol-friendly mechanism inside the Protocol Translator that can increasethe conversion speed. The key point of this mechanism is a protocol Name-Valueindex table of data which is carried in the optional headers of the differentapplication protocols. In TCP/IP protocol suite contains at different levelsdifferent application protocols, some of the important protocols are CoAP, REST, MQTTM, MQTT-SN, AMQP.
When apacket reach at the gateway, the Protocol Translator examines the optionalheader. If it determines the index table there, then it grabs the dataimmediately from the payload it place the packets in destination protocol. Inindex table is stored as on optional header, application protocols may not use the index tables. In such cases, theconversion is done in the conventional form and consequently it takes span oftime.
Figure 5: (a) Optional header of the application protocol and Index table (b) The conversionmechanism inside the gateway. In Figure 5, Inoptional header there are variousapplication protocol formats that assigned index number in index table. In index table wherea packet consisting of Name-Value pairs suppose for example x-97, y-99 etc. they needed to be converted in gate way from asource protocol to a desired protocol in the protocol translator format. The data is storedin a linear structure inside the payload of each packet.The analysis ofprotocol translator has been implemented XMPP, in which data are stored in XMLtags. In application layer format, it need O(n/2)operations are required to find a data item in the payload before inserting it into thedesired protocol.
There are O(n2) operations arerequired in name-value pairs data inside packets. The conversion of XMPP takesthe if the position of each Name-value item is available then the conversion time will be reduced to O(n). V. Conclusion Inthis paper outlined the vision of Big Data Analytics, Fog and Cloud computingtheir role in IoT and future of IoT.
The layered architecture of IoT systemperformed by the IoT framework and IoT elements. The IoT architecture of thismassive infrastructure has been defined, store and network devices. thesimulation of protocol translator provided by application layer in IoT isdefined. In future there is scope to investigate new Big data analyticsplatform, Fog and Cloud computing platform of IoT, deliver a portfolio of newservices in IoT and develop better and services of IoT.