# Introduction he Internet of Thing (IoT) refers to gathering and sensing data from various apparatus about our daily life corporal phenomena [1,2]. IoT has acquired capacious popularity owing to its extensive market in several sectors such as home automation, healthcare, security, etc. [3]. IoT entangles pervading internet connectivity beyond valuable devices to any province of traditionally daft or non-internet-enabled naturalistic devices and quotidian things [4]. IoT was begotten from machine-to-machine (M2M) interaction. Machines are joining to the objects via a network except for human communication [5]. Real-time is a level of computer compassionate that a customer senses as sufficiently early or that qualify the computer to repose up with several processes (for example, to attendant visualizations of the weather as it incessantly mutations) [6]. Real-time is an epithet regarding computers or processes that handle in real-time. Real-time describes a man rather than an engine sense of time [7]. Real-time computing (RTC) is a period of computing decision that has inelastic instant constraints. Real-time computing must be committed in a time structure that is relatively invisible to the user [8]. A real-time system is any message processing system which must answer to externally propagated input incentive between a limited and specified epoch [9]. If it overcomes this time bondages, it's output in accomplishment erosion and malfunction of the process [10]. Denoting to the data-scheming system in which a computer takes every moment transforming data, such as knowledge relating to air-traffic monitoring, travel booking procedure and processes it practically to be capable of controlling the origin of data [11]. # a) Characteristics of Real-Time System ? Time Encirclement. ? Modern Rectification Inference. ? Interwoven. ? Security-Criticality. ? Concurrency. ? Allotted and Recompose Formation. ? Ought Troublesome. ? Usage Hardware. ? Durability. # II. # Background and Related Work In 1970s Minicomputers specially built for dedicating embedded systems such as DOG scanners, onward the necessity for low-latency priority-driven with accessing data and operating systems suchlike Data General's RDOS (Real-Time Disk Operating System) and RTOS's background and foreground scheduling as Digital Equipment Corporation's RT-11 date from this age. On time when the MOS Technology 6502 and the Motorola 68000 were exoteric relatives could conduct their abode computer as a real-time system. The chance to deactivate several intermissions allowed for hardcoded loops with fixed timing and the less interrupt latency allowed the impersonation of a real-time system [12]. We implemented and modified our system with this two-existing system. In the procedure 1, Smart home regulating devices are using for guiding the systems by sending data to govern the actuators. Here they also used the home gateway for relating all home appliances and interconnected with the Internet on the open network part, and the home networks also control the home machinery. The home gateway gives the ability to control the connected equipment on the home networks that controlled with smart instruments and passed data through the system and collect data for calculating real-time [13]. In the procedure 2, there submitted a low cost secure mobile phone-based home automation system [14,15]. Appliances at home are linking with the Arduino BT board, and the communication within the mobile phone and the Arduino BT board is wireless [16]. Superfluous apparatus can be connected into the system, and it should be faster enough to actualize the power of wireless technology and be cost effective. In this system, they also are measuring data for calculating real-time [17,18]. In our system, we connected all devices with the server and motion sensor for controlling our automation system. We passed data through our implemented process for measuring the time that is needed to turn on/off the equipment. For calculating real-time performance, we implement a virtual system in Cisco Packet Tracer. We use server and motion sensor for connecting the devices. The equipments are light, fan, door, window, sprinkler, webcam and motion detector. The sensor is joining to the SBC-PT Board, which is one kind of networking access board. All equipment attached with the server through a unique IP address. Entering to the server and viewing the activity of the system by home authority there use a laptop. To collect real-time data, we passed packets through the system and putting it to the performance matrices equation for calculating the real-time action of our system. We also build the other two existing methods in Cisco Packet Tracer for comparing our systems performances [19]. # III. # Proposed System a) System Description # b) User Case Diagram User case provides a path, how the user gets the entrance license of the system. A user cannot straightly access the arsenal so that they needed to exact authentication. Here also we see equipment internally linked with the system action. Use cases are typically induced by a user to meet the goals detail activities and variants inlaid in gaining the aim [20]. # Flow: ? Admin Managing & Controlling Home Appliances Logout. Admin is the prime controller for managing, measuring and controlling the access of the services. # Fig. 2: User Case Diagram ? User Login Using Equipment Logout. To start using the system, the user must use the login interface to log in into the system. When the user wants to control the equipment, they first log in to the service with their password. If their passwords match, then they get the access. After using the device, they are log Out from the process. ? Ingredient Motion Sensor. Light, Fan, Door, Window, Sprinkle connected to the motion sensor. When the sensor detected any person, then it is automatically on/off. ? Ingredient Motion Detector. Webcam connected to the motion detector. If any person passes through a motion detector, then it is automatically on/off. Therefore, we see the passing passenger. # c) Layer Development Layering is the association of programming into individual functional elements that interact in various sequential and hierarchical way, including each layer generally having an interface alone to the layer above and below it [21]. IV. # Analysis Of Proposed System a) Data Collection For calculating real-time, we pass packet through equipment from starting point to end. We pass the packet into two different ways. And the ways are 1) Equipment to server & 2) Server to Equipment. In this process we listed time, some packets are passed that called passing time or true position time. Here true position time denoted as x & y. And some packets are not passed that called obstacle time or false position. We denote false position as x' & y' [22,23]. To calculate real-time, we use six different types of performance metrics. The metrics are Sensitivity, Specificity, Precision, Accuracy, F1-score, G-mean. We can apply all the performance metrics into seven different equipment such as light, fan, door, window, sprinkler, webcam and motion detector. ? G-mean: The G-mean is computing by taking the square average value of sensitivity and specificity. G-mean =?Sensitivity + Specificity # b) Comparison Previously we discussed two different systems. We can also implement these systems and passes packet through that systems for calculating the passing time and obstacle time. With the help of performance metrics, we calculate real-time through this passing time and obstacle time. The Graph for Light (Sensitivity): System 1 goes zigzag, and system 2 goes from high to low but our system goes from low to high. The Graph for Light (Specificity): Our system's specification is better than other two process because it is an up-ordering line. The Graph for Light (Precision): # Comparison of Light: Our system precision is better than other. The Graph for Light (Accuracy): System 1 and system 2 are creating down order line, but our system is high-ordered. The Graph for Light (F1-score): The value of F1-score in our system is excellent than the other two methods. The Graph for Light (G-mean): We are measuring the value of three systems and plotting the value for generating the graph line, and line is increasing. The Graph for Fan (Sensitivity): For comparing three systems, we saw that our system's sensitivity is better than others. The Graph for Fan (Specificity): We collect data and compare with them and saw that the proposed system is proper. # The Graph for Fan (Precision): From the table and graph, we predicted that the precision of our system is the best than others. The Graph for Fan (Accuracy): System 1 and 2 is down order line. Our system is emersion, so it is enriching the graph. The Graph for Fan (F1-score): Plotting the value and choose a better system and our system is better than others. The Graph for Fan (G-mean): For fan, the G-mean is elder than system 1 and system 2 because it is emergence from others. The Graph for Door (Sensitivity): System 1 rise low to high and low. System 2 is ascent high to low and high. Our system goes low to high without breaking. The Graph for Door (Specificity): Plotting the value and generate the graph for three system and our system graph is better. The Graph for Door (Precision): For metering the effect of the specification of fan we collect and compare data for three system and the proposed system is strong. The Graph for Door (Accuracy): System 1 and 2 are devious, and our system's accuracy is the best because it is a rising line. The Graph for Door (F1-score): Construct the diagram we saw that system 1 and system 2 are decreasing, and our system is increasing. The Graph for Door (G-mean): Our system's G-mean is elder than system 1 and system 2 because it is emergence from others. From the diagram, we saw that our proposed system sensitivity is good than other systems. The Graph for Window (Specificity): For monitoring the effect of specificity, we collect data and compare them and choose the better line. The Graph for Window (Accuracy): Our system is heightening from system 1 and system 2 because it is a prosperous order line graph. From the table, we plot the value and choose a better system, and the implemented system is the best than others. The Graph for Window (G-mean): We are measuring and plotting the value for the diagram and observe that it is increasing. In modern time almost, people prefer to do their work digitally, so they use the internet to do that work. Through the internet, we connect with those equipments easily which we used in our daily life for controlling that thing. With the help of IoT and the Internet, we can do those activity [24]. We can control and monitor our home equipment when the owner goes outside. For this purpose, this system needs to response in less time because if the materials turn on a long time, it will damage, or its cases more electric bill. We try to remove that problem in our paper. In future, we develop this system for consuming time on demand [25,26]. V. # Result In this system, we calculate the time for seven equipment, and here we give four equipment's (light, fan, door, window) real-time in-home automation system. In our next paper, we will discuss and calculate the time for other three equipment. We passed data packet through every device for ten different ways. Collecting values and plotting those for creating a graph that helps us to view real-time response. We excrement for measuring a mediocre period and comparing with other existing systems. From the diagram, we see that our systems Sensitivity, Specificity, Precision, Accuracy, F1-score, and G-mean provide a better graph line than others. And we see that our systems response is better than another system. # VI. # Conclusion IoT promotes the proficiency of existence by joining the digital world to the real world via setting different sensors in daily life objects [27]. If we utilize our system correctly, we save our valuable time, energy; electric power moreover comforts of life. This work may be expanded to additional provision from other networks as those rules may provide better insight into the conduct [28]. 1![Fig. 1: System Implementation of the Proposed System](image-2.png "Fig. 1 :") 3![Fig. 3: Layer Development ? Physical Layer: It gives the hardware using sending and receiving data on a carrier, together with demarking cables, cards, and physical aspects. ? Data Link Layer: This layer shows the functional and procedural means to trek data between network essences and detects errors that occur in the physical layer. ? Network & Transport Layer: The network layer is consisting of a variant form of wired and wireless contact networks such as Wireless Local Area Network (WLAN), Internet. This layer is liable for the sending and receiving information from the perceptive layer. Transport layer is a conceptual partition of a scheme in the layered masonry of protocols in the network pile in the Internet Protocol Suite. The transport layer is the foundation of the Internet. ? Presentation Layer: It offers protocol conversion, data translation, compression, and encryption. It provides the real-time succession requisite for communication between objects that are layers, systems or networks. ? Application Layer: This layer creates an interface between the user and system performance.](image-3.png "Fig. 3 :") ![Sensitivity: Sensitivity contents the proportion of true positives that are identified. Sensitivity = ?? ??+??? ? Specificity: Specificity extents the proportion of real negatives that are identified. Accuracy: Accuracy is the caliber of the instrument to the volume the actual value. It is a tracing of systematic fault, a measure of statistical favor. Accuracy = ??+?? ??+?? ? +??+???? F1-score: The F1-score (F1-sc.) is the harmonic mean of precision (Prec.) and sensitivity.](image-4.png "??F1") 4![Fig. 4: Light (Sensitivity)](image-5.png "Fig. 4 :") 5![Fig. 5: Light (Specificity)](image-6.png "Fig. 5 :") 6![Fig. 6: Light (Precision)](image-7.png "Fig. 6 :") 7![Fig. 7: Light (Accuracy)](image-8.png "Fig. 7 :") 8![Fig. 8: Light (F1-score)](image-9.png "Fig. 8 :") 9![Fig. 9: Light (G-mean) Comparison of Fan:](image-10.png "Fig. 9 :") 10![Fig. 10: Fan (Sensitivity)](image-11.png "Fig. 10 :") 11![Fig. 11: Fan (Specificity)](image-12.png "Fig. 11 :") 12![Fig. 12: Fan (Precision)](image-13.png "Fig. 12 :") 13![Fig. 13: Fan (Accuracy)](image-14.png "Fig. 13 :") 14![Fig. 14: Fan (F1-score)](image-15.png "Fig. 14 :") 15![Fig. 15: Fan (G-mean) Comparison of Door:](image-16.png "Fig. 15 :") 16![Fig. 16: Door (Sensitivity)](image-17.png "Fig. 16 :") 17![Fig. 17: Door (Specificity)](image-18.png "Fig. 17 :") 18![Fig. 18: Door (Precision)](image-19.png "Fig. 18 :") 19![Fig. 19: Door (Accuracy)](image-20.png "Fig. 19 :") 20![Fig. 20: Door (F1-score)](image-21.png "Fig. 20 :") 21![Fig. 21: Door (G-mean)](image-22.png "Fig. 21 :") 22![Fig. 22: Window (Sensitivity)](image-23.png "Fig. 22 :") 23![Fig. 23: Window (Specificity)](image-24.png "Fig. 23 :") 24![Fig. 24: Window (Precision)](image-25.png "Fig. 24 :") 25![Fig. 25: Window (Accuracy)](image-26.png "Fig. 25 :") 26![Fig. 26: Window (F1-score)](image-27.png "Fig. 26 :") 27![Fig. 27: Light (Sensitivity) c) Future WorkIn modern time almost, people prefer to do their work digitally, so they use the internet to do that work. Through the internet, we connect with those equipments easily which we used in our daily life for controlling that thing. With the help of IoT and the Internet, we can do those activity[24]. We can control and monitor our home equipment when the owner goes outside. For this purpose, this system needs to response in less time because if the materials turn on a long time, it will damage, or its cases more electric bill. We try to remove that problem in our paper. In future, we develop this system for consuming time on demand[25,26].](image-28.png "Fig. 27 :") 1Equipment toServer toServerEquipmentEquipment to ServerPassing Time(x)Obstacle Time(x')Server to EquipmentObstacle Time(y')Passing Time(y) 2LightSystem 1System 2Proposed System0.3390.7590.3680.2730.7400.3980.4120.7660.4610.3490.7830.484Sensitivity0.402 0.3530.793 0.7990.518 0.5710.4140.7860.6710.3740.7760.6690.4140.7820.6790.4520.7870.805 3LightSystem 1System 2Proposed System0.4670.6920.3940.2930.7050.4210.3880.7890.5190.4460.7100.538Specificity0.493 0.4570.719 0.7120.541 0.5970.4750.6970.6690.5030.6870.6810.5250.6680.7210.5030.6730.881 4LightSystem 1System 2Proposed System0.4370.7430.4790.2660.7270.5030.3730.7610.5300.3320.7200.532Precision0.409 0.3770.727 0.7180.539 0.5540.4110.7030.5580.3950.6920.5750.4350.6730.6220.4590.6770.708 5LightSystem 1System 2Proposed System0.3970.7290.4990.2840.7230.5000.3990.7580.5280.4030.7460.537Accuracy0.450 0.4070.755 0.7540.551 0.5760.4470.7390.5850.4430.7290.6250.4730.7210.6690.4780.7260.821 6LightSystem 1System 2Proposed System0.3820.7510.4170.2690.7340.4560.3910.7640.4930.3400.7500.517F1-score0.406 0.3650.758 0.7570.529 0.5730.4130.7420.5750.3840.7310.6090.4240.7230.6930.4550.7280.702 7LightSystem 1System 2Proposed System0.8981.2050.9820.7521.2021.0010.8941.2471.0280.8921.2221.033G-mean0.946 0.9001.229 1.2291.044 1.0440.9431.2181.0690.9361.2091.1260.9691.2041.1600.9771.2081.249 8FanSystem 1System 2Proposed System0.4970.4250.2180.5240.7980.2320.5580.8250.2430.5860.8670.248Sensitivity0.579 0.5760.779 0.7910.252 0.2550.5660.7620.2640.5700.7750.2830.5690.7520.3750.5550.7640.454 9FanSystem 1System 2Proposed System0.3230.8540.2920.4910.7520.3340.5950.7830.3940.4960.7360.437Specificity0.523 0.5710.767 0.7620.499 0.5270.5790.7770.5720.5940.7870.5930.5730.7650.6480.5720.7740.843 10FanSystem 1System 2Proposed System0.5450.3920.2710.5820.6880.2800.5520.7520.2870.5380.7770.292Precision0.555 0.5970.745 0.7890.305 0.3480.6010.7600.3570.6130.7740.3670.5770.7510.3890.5630.7630.513 11FanSystem 1System 2Proposed System0.4310.7770.2680.5090.7710.2940.5280.8020.3150.5410.8040.355Accuracy0.552 0.5730.773 0.7780.374 0.3890.5720.7690.4530.5810.7810.4890.5710.7580.5030.5640.7690.739 12FanSystem 1System 2Proposed System0.5200.4080.2420.5510.7390.2540.5550.7870.2630.5610.8190.271F1-score0.567 0.5860.762 0.7910.274 0.2970.5830.7610.3000.5910.7740.3190.5740.7510.3820.5590.7640.481 13FanSystem 1System 2Proposed System0.9061.1310.7351.0071.2450.7731.0261.2680.8061.0401.2660.849G-mean1.049 1.0711.243 1.2460.852 0.8631.0701.2410.9141.0791.2490.9491.0691.2320.9731.0621.2401.139 14DoorSystem 1System 2Proposed System0.4550.6380.0960.5220.5420.2360.6070.4560.2720.6120.6280.289Sensitivity0.615 0.5780.704 0.7390.298 0.3040.5550.7610.3090.5720.7720.3290.5520.7810.3570.5260.7880.367 15DoorSystem 1System 2Proposed System0.4270.7870.4860.4460.8200.5540.4860.8260.5730.4630.7890.615Specificity0.492 0.4950.773 0.8050.619 0.6440.5360.7590.6490.5230.6870.6660.5200.7620.7370.5420.7470.949 16DoorSystem 1System 2Proposed System0.4920.7560.3340.4670.5890.3340.4980.4860.3350.4730.5880.335Precision0.499 0.5000.641 0.7170.336 0.3380.5400.7160.3470.5270.6460.3650.5060.7320.3730.5080.7450.391 17DoorSystem 1System 2Proposed System0.4430.7130.3990.4820.7310.4380.5410.7280.4650.5290.7370.494Accuracy0.547 0.5340.748 0.7790.496 0.5150.5450.7600.5420.5470.7230.5480.5350.7710.5550.5340.7670.751 18DoorSystem 1System 2Proposed System0.9830.6880.1520.4930.5650.2780.5470.4700.2810.5330.6070.310F1-score0.551 0.5360.671 0.7280.315 0.3190.5480.7380.3210.5490.7030.3380.5280.7560.3630.5160.7660.378 19DoorSystem 1System 2Proposed System0.9391.1940.8850.9841.1670.9181.0451.1320.9391.0361.1900.957G-mean1.052 1.0351.215 1.2430.959 0.9861.0451.2330.9891.0461.2080.9971.0351.2421.0081.0331.2391.022 20WindowSystem 1System 2Proposed System0.5070.3180.1500.5540.8050.2440.6010.8180.2540.6220.7570.256Sensitivity0.622 0.5870.781 0.7420.283 0.3030.5840.7630.3170.5870.7350.3180.5840.7530.3220.5810.7310.325The Graph for Window (Sensitivity): 21WindowSystem 1System 2Proposed System0.5940.1870.5490.6470.0350.5700.6140.5410.6170.6360.4990.621Specificity0.576 0.6070.655 0.6150.622 0.6490.5990.6730.6620.6110.6450.6770.5980.6610.7160.1410.6390.969 22WindowSystem 1System 2Proposed System0.6010.2760.2490.6500.6900.2580.6160.7020.2590.6380.6660.288Precision0.577 0.6080.740 0.7060.325 0.3340.6000.7340.3340.6120.7080.3340.5990.7090.3340.6070.6890.339The Graph for Window (Precision):System 1 and 2 create sprawl line. Our system makes a straight line. It is favorable. 23WindowSystem 1System 2Proposed System0.2090.2520.4680.5970.5950.4980.6070.6990.5020.6290.6460.507Accuracy0.598 0.5970.725 0.6860.519 0.5460.5910.7220.5470.5990.6940.5480.5910.7090.5850.4470.6870.894 24WindowSystem 1System 2Proposed System0.5490.2960.2470.5980.7430.2540.6080.7550.2560.6290.7090.257F1-score0.599 0.5960.399 0.7230.286 0.3180.5920.7480.3210.5990.7210.3250.5910.7300.3280.5940.7090.329The Graph for Window (F1-score): 25WindowSystem 1System 2Proposed System1.0490.7110.9351.0950.9170.9371.1021.1660.9421.1221.1210.945G-mean1.095 1.0921.198 1.1650.951 0.9651.0881.1980.9691.0951.1750.9891.0871.1891.0091.8491.1701.058 © 2019 Global JournalsAn Analytic Real-Time Framework for IoT Based Home Automation System * Research Directions for the Internet of Things JAStankovic IEEE Internet of Things Journal 1 Feb. 2014 * Internet of Things (IoT): A vision, architectural elements, and future directions JGubbi RBuyya SMarusic MPalaniswami Future Generation Computer Systems 9 7 September 2013 Pages 1645-1660 * HOME AUTOMATION USING INTERNET OF THINGS ABhat Sharma Pranav IRJET) Volume: 04 Issue: 07 International Research Journal of Engineering and Technology July -2017 * IOT and Cloud Computing in Automation of Assembly Modeling Systems CWang ZBi LiDaXu IEEE Transactions on Industrial Informatics 2 10 May 2014 * The Internet of Things vision: Key features, applications and open issues EBorgia Computer Communications 54 1 December 2014 * Smart Home-Control and Monitoring System using Smart Phone RPiyare SRLee ICCA, ASTL, 2013 -researchgate.net * Real-Time Systems KJuvva Spring 1998. April-2018 Online; accessed 11 * IoT based monitoring and control system for home automation DPavithra RBalakrishnan Global Conference on Communication Technologies (GCCT) 2015. April 2015 * An Efficient Algorithm for Mediabased Surveillance System (EAMSuS) in IoT Smart City Framework VAMemos KEPsannis YIshibashi BGKim BGupta Future Generation Computer Systems 25 April 2017 * IoT based Monitoring and Control System for Home Automation MAshok PSVarma MV RSundari International Journal of Engineering Technology Science and Research IJETSR * Major Requirements for Building Smart Homes in Smart Cities based on Internet of Things Technologies TK LHui RSSherratt DDS´anchez October 3. 2015. 2010. April-2018 2018 PubNub Inc 58369 Real-time computing. Online; accessed 22 * Smart Home System .RMr Kadam .PMr Mahamuni Mr Parikh International Journal of Innovative Research in Advanced Engineering (IJIRAE) 2349-2163 2 1 January 2015 * Home Automation Using Internet of Things PPatel MPatel VPanchal & VNirmal Imperial Journal of Interdisciplinary Research (IJIR) 2 5 2016 * Mobile based home automation using Internet of Things (IoT KMandula RParupalli ChA SMurty EMagesh RLunagariya 2015 International Conference on Control, Instrumentation, Communication and Computational Technologies (ICCICCT) 19 Dec. 2015 18 * Internet of Things: Ubiquitous Home Control and Monitoring System using Android based Smart Phone RPiyare International Journal of Internet of Things 2013 * Bltooth Based Home Automation System Using Cell Phone RPiyare MTazil 2011 IEEE 15th International Symposium on Consumer Electronics June 2011 * MavHome: An Agent-Based Smart Home DJCook MYoungblood EOHeierman KGopalratnam SRao ALitvin FKhawaja Proceedings of the First IEEE International Conference on Pervasive Computing and Communications the First IEEE International Conference on Pervasive Computing and Communications 26 March 2003 * A Security Framework for IOT based Smart Home Automation System NHossain MAHossain RSultana FALima Global Journal of Computer Science and Technology (GJCST) 18 0 Jun. 2018. 2005 Certified Group * Smart Home: Integrating Internet of Things with Web Services and Cloud Computing MSoliman T. Abiodun THamouda JZhou CHLung IEEE 5th International Conference on Cloud Computing Technology and Science Bristol, United Kingdom United Kingdom 2013. 2013. Dec. 2, 2013 to Dec. 5, 2013 * Study on Smart Home System Based on Internet of Things Technology YLiu Informatics and Management Science IV Lecture Notes in Electrical Engineering WDu London Springer 06 December 2012 207 * Fog Intelligence for Real-Time IoT Sensor Data Analytics HMRaafat MSHossain EEssa SElmougy ASTolba GMuhammad AGhoneim IEEE Access 5 20 September 2017 * Chance Constrained Optimization in a Home Energy Management System YHuang LWang WGuo QKang QWu IEEE Transactions on Smart Grid 9 1 Jan. 2018 * Home Automation Using Internet of Things PPatel MPatel VPanchal V Imperial Journal of Interdisciplinary Research (IJIR) 2 5 2016 * Internet of Things based Home Automation System SSoumya MChavali SGupta NRao 2016 IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology (RTEICT) Bangalore, India May 2016 * A review of smart homes-Present state and future challenges MChan CEstève Escriba Campo * Computer Methods and Programs in Biomedicine 91 1 July 2008 * Indoor air quality and energy management through real-time sensing in commercial buildings PKumar CMartani LMorawska LNorford RChoudhary MBell MLeach Energy and Buildings 111 1 January 2016 * Secure integration of IoT and Cloud Computing CStergiou KEPsannis BGKim BGupta Future Generation Computer Systems 1 December 2016