Adaptive And Fault-Tolerant Data Processing In Cloud Environment: Review
G. Jasmine Christabel
1, A.C. Subhajini
21,2Department of Computer Application, Noorul Islam Centre for Higher Education, Thuckalay, Kanyakumari, India, 629180.
E.mail: acsubhajini@gmail.com, jasminejcs@gmail.com
Article History: Received: 10 January 2021; Revised: 12 February 2021; Accepted: 27 March 2021; Published
online: 28 April 2021
Abstract: Cloud technology is a modern technique that utilises assets at the edge of both the cloud purpose of
providing resources for IoT (Internet of Things) system requests. The highly distributed network design of cloud network edge allows faults to be a common event and condition monitoring has therefore become important. Time-sensitive demands were not strongly regarded by most initial planning and fault-tolerant techniques. This raises the likelihood of latencies that trigger unfavorable results in fulfilling certain demands. The key purpose is to minimize service latency and running costs and to improve the cloud's stability and power.Undoubtedly, the dynamism and strength of the cloud and its services is the most prevailing topic of this entire decade. However, there come huge challenges with this ease of dynamic cloud environments. Cloud is prone to faults and errors; different services behave differently and can generate a series of unexpected errors. Here comes the system ability to handle these errors gracefully, moreover, the actual challenge is to make this handling effective in the long run. The primary concern for maintaining service stability and programme execution of virtual servers is fairly flexible. Failures must be expected and constructively managed to mitigate the failure effect of the virtualization and programme execution.To foresee these errors and undertake sufficient steps before faults start happening, fault tolerance strategies have been used. This report attempts to provide a deeper understanding of the complexities of appropriate fault-tolerant and describes different methods and strategies used during resistance of faults.
Keywords: Cloud technology, virtual machineries, fault tolerance, data processing Introduction
In context of cloud applications, cloud technology refers to the access, configuration and manipulation of assets like software and hardware at such a distant location specified by cloud services. A cloud is a set of parallel-baseddisseminated scheme comprising a collection of continuously protection of environment integrated and virtualization technologies machines which are represented also as or even more centralized computing power focused on service-level agreements formed via negotiations amongst service providers and customers [1]. Cloud computing has led to a shift from a commodity to a service mostly in distribution view of information technology. It has made it possible for different applications, networks and infrastructure development tools to be made available as substantial return on investment on request and over internet [2]. Nevertheless, relation to the large where they run, the efficiency of cloud services is hindered owing to its inherent susceptibility to errors. Unless the performance-linked problems of reliability, obtainability, and efficiency were done properly via cloud storage workers [3,4] will cloud computing services be utilized for the maximum capabilities. To accomplish higher efficiency in cloud services, fault tolerable is thus a vital prerequisite. This article presents a detailed overview of cloud services concerns associated to fault resistance; stressing the relevant principles, state-of-the-art tools and procedures. The aim would be to provide overview of the current strategies to fault resistance as well as problems which have yet to be solved.
Another of the new technologies in an arena of informative systems is cloud computing [5]. It includes enormous resources of computation which can be obtained via a compensation-per-use programme. For cloud technology, virtual machine plays a key role [6,7]. It simplifies the control, via a hypervisor named Virtual Machine Monitor (VMM), of any and all assets as a singular entity. It can include software and services (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS) focus on best processing capacity, efficient processor and useful shared resources principles. It is variable in nature, usually by both the minutes and the hour, and can sale on requests. The Cloud Service Provider (CSP) completely operates the product. Excluding creating new infrastructure, educating new employees or licensing new technologies, this is a way to expand power or add functionality and enhance speed comparison to supercomputers. Virtualization technologies, edge systems and a poor economy are of serious importance to cloud technology.
Servers to store information at any time that use the applications programming interface (API) so that cloud services can view it from any Internet-connected terminal device [8]. The power stations offered by hardware / software
facilities are not always open for public voting and were accessible in the corporate market. Service provider’s offerings could be anything from the infrastructure, network, and software tools. Each one of these resources, as seen in Figure 1, is for platform as a service (IaaS), Platform as a Service (PaaS) or Software as a Service (SaaS). Hence, fault tolerance and scheduling are two important parameters for reliable system nodes. Due to the vast complexities of cloud, to deliver a comprehensive resolution for cloud safety will be difficult. Cloud computing environment in two basic characteristics timeliness and fault tolerance is able to work. With timeliness, we mean that any work in real time, a time limit is that Run it will end and fault tolerance means that work must continue to error. Cloud support is really important for real-time systems [9,10]. In general, a real-time system is an information processing system in response to external stimuli input production, within aninadequate and specified range of time. In cloud computing, the use of processor nodes, and also because it raises the probability of error in terms of safety-critical real-time systems should increase their reliability and hence the demand for achieving fault tolerant systems in real-time systems to enhance [11].
Fig.1. Types of cloud computing
Cloud computing is used for retrieving, organizing and operating the software and hardware sources from a remote location [12]. It tenders application, infrastructure, and online data storage. Cloud computing communicates which we can access the applications as utilities over the internet. It is used to customize, configure, and create business applications online. Burden tolerance is a substantial problem to assurancea reliability and availability of seriousfacilities and also execution and application processing [13]. Based on the disappointmentinfluence on an application execution and system, the failure mode must be handled predicted and pro-active in nature. Erroracceptance methods were objective to predetermine thisdisappointments and initiate proper exploit before letdownsessentiallyhappen. In thesearticle, explain an overviews of the current burdenbroad-mindedness methods in cloud technologyowing to itsstrategies, research challenges, and techniques utilized.
Cloud technology is a model which enables fast, through the-demand channel access to a common pool of configurable, such as computers, channels, storage devices, services, and uses, which could be rapidly provisioned and distributed with the least project implementation or interaction between service providers [14,15]. Cloud technology provides multiple on-demand tools for end customers in a form of services. It allows companies and consumers to utilize apps in the absence of installing of physical devices and permitsentree over the Internet to a necessary services. This offers characteristics such as higher performance, pay-as-you-go, accessibility, interactive content, consistency, fully programmable ease, productivity, scalability, broad information management and permeability to turn IT from a commodity to a service [16].
As a rapidly evolving technology, cloud computing is widely sometimes used host several related to the business systems. The widespread need for cloud-based platforms for corporation or business hosting purposes, nevertheless, refers to trouble with customer satisfaction and usability for both providers and users [17]. And for its large scalable nature, asset complexity and the large scale of activity, these concerns are inherent to cloud services. Therefore, in the production environment, many types of errors can occur, leading towards errors and quality degradation [18,19]. The massive flaw kinds are classified as follows:
a) System failure: Because cloud computing services are accessible over even a network (Internet), network errors are now the main causes of cloud services losses. Such faults can happen due towards node failures, packets data corruption, fragmentation, destination address or connection failure, etc.
b) Physical defects: These are flaws that exist predominantly in underlying hardware, like CPU, memory, storage, power outage, etc.
c) System defects: weaknesses in procedures can occur due to the shortage of assets, software bugs, inadequate process technology, etc.
d) Service expiration fault: If the time served of an asset expired when it is being used by an operation that rented it, this contributes to system failure.
Faults contribute to a malfunction or termination of a process. Cloud computing, and therefore data storage, is, nevertheless, distinguished by the principle of incomplete failures [20]. In either constituency device, method or schedule, a fault can arise. This contributes to a system failures and, therefore, rather than a full collapse, performance deterioration. While this resulted in stable and efficient systems, the correct fault - tolerant frameworks for higher-performance computing can better manage faults. Fault tolerance allows the application to be supported by the device even though few of its components do not function properly.
Fault tolerance (FT) was its capacity of a process which, irrespective of faults, continues to play its expected function [21]. In other phrases, FT is linked to precision, efficient operation, and lack of disruptions.A FT-depended solution must be designed to investigate and somehow still meet any requirements for errors in specific software or hardware elements, power losses or another variations of unforeseen struggles.
Circumstantial of cloud technology
Cloud storage has developed from initiatives in multiple computational research fields like centralized computing, grid computing, technologies of enterprise applications and Service-Oriented Architecture (SOA). These have, nevertheless, often imbibed the traits, innovations, and limits. An infrastructure of cloud technology involves the machines, memory devices, network facilities and other associated systems needed to provide users with cloud services tools and services [22,54]. These elements of hardware are often situated inside corporate data centres. These include multi-core computers, solid-state drives and hard disks, all on a wide scale, providing storage facilities and networking gear, like firewalls, switches, and routers. Software modules that enable the cloud storage model, like virtual servers, are often referred to as the cloud computing platform, besides these hardware components [23]. Virtualization technique offers an overview to cloud services and uses Application Program Interfaces (APIs) and other command - line interface and/or interactive interfaces to typically give users these tools. Virtualization services are typically distributed to users through the internet (and often over any other network) provided by cloud service providers (CSPs).
In general, cloud processing capabilities are delivered to users as resources on a distributed and standalone treatment-based strategy. Big CSPs, such as Amazon Web Services (AWS) and/or Google Cloud Platform [24], choose an axillary region-based strategy. Such method is being utilized to work collaboratively between various uses and tenants (businesses, organizations, etc.) employing the cloud in a cost-effective but safe manner. In order to make sure separation amongst tenants, application virtualization can be utilized [55]. Clients, servers, software, as well as another elements form a standard cloud platform. The distributed system (DFS), such as the Google File System (GFS) and/or Hadoop Distributed File System (HDFS), is yet another cloud storage part that is primarily employed to store information on discs throughout the formation of objects or blocks [25]. Such data structures disentangle the transfer of records from of the real physiological data and therefore guarantee storage usability. Cloud computing environment thus comprises, in general, of cloud computing platform.
a. Servers-the actual devices for one or maybe more virtual machines which act like physical machine. b. Cloud computing-Technology which abstracts and offers as logical assets hardware elements like databases, storaging, and networking.
c. Storage: Network connected storage (NAS), disc drives etc. and in context of Storage Area Networks (SAN). Encrypting and recovery, alongside equipment.
d. Network- system provides virtual data and storage interrelationships.
e. Management-Different tools for network infrastructure setup, coordination and monitoring, like servers, networks, and hard drives.
f. Protection-Components that, in particular, include data security, availability, or confidentiality and information assurance.
g. Tools for backup and restore.
Cloud dispositionprototypes
i. The motivation and setting where a community cloud is supposed for use is the basis of the cloud solution. The option of the service models defines the expenses involved, the resource usage of energy and other
construction expenditures [26,27]. In public clouds, the most widely used application models are public and private cloud, group cloud, and hybrid cloud.
ii. i. Public Cloud: It offers access for the wider populace to the networks and applications provided by an organization supplier [56]. It gives a really less cost of versatility, scalability, promoting independent as multi-tenancy is commonly used. With the on-demand basis, services are continuously delivered by a centralized third-party provider which offers financial and use a multi-tenant strategy.
iii. ii. Private Cloud: It is used by a single entity, i.e. within an enterprise, cloud services and goods may be obtained or used [57].This method guarantees optimum privacy and security of the configuration information.
iii. Community Cloud: This framework is used concurrently by multiple companies / organisations and assists a specific community / social system that requires community engagement (such as security needs, mission and enforcement concerns, etc.)[58]. One or more organizations within the group or / and a seller can run, own, and maintain this template.
iv. Hybrid Cloud: This contains both a private cloud and a cloud infrastructure partnership. Overall accountability (e.g. those involving stable operations) are achieved utilizing private cloud resources in this cloud implementation, and non-critical activities are managed to use the public cloud[59].
iv. In situations where governments choose to utilize communication services such as chat and video meetings, cloud systems are most acceptable, but adequate IT assets or facilities are not centrally available [28]. In addition, a proprietary deployment period could be used if stringent protection and confidentiality are concerns of higher priority. On another hand, a mixed cloud platform ought to be the alternative for an organization that has a broad IT framework but is also growing its capacities.
v. Cloud service prototype
While cloud storage has developed dramatically in recent years, conventional attendance in three major business structures [29]. The basic service modelsare
Software-as-a-Service (SaaS): throughout this model, the cloud service provider presents software programs in the context of consumer / end-user products. An application supplied to the customer as a service eliminates such a need to install and conduct the public cloud mostly on computer of the customer and facilitates management. For instance, web unified communications services, email systems, opportunities for online media, etc. Amazon AWS, google cloud platform, Microsoft Azure, IBMM, and the catalogue of cloud services are SmartCloud Enterprise, Cloudand OpenStack, Open-Nebula, Qstack and etc. [60].
Platform-as-a-Service (Paas): This framework offers an opportunity for both the cloud creation, implementation, testing and management of technologies. With a database engine from a CSP, a consumer can rent an atmosphere and use that for custom uses creation. Acquia Cloud, App Agile, Apprenda, Bluemix, Cloud 66, Cloudways and etc are the collection of PaaS providers [61]. Infrastructure-as-a-Service (IaaS): The IaaS Company provides links to certain main source, such as physical computers, processing, networks, servers, cloud-based virtual computers, etc. Services including flexible virtual server providing or on-demand warehouses are offered by the IaaS provider. Sales Force, Microsoft, Amazon Web Services, and the number of SaaS providers are Slack, Zendesk, Oracle, GitHub, Cisco etc [62].
Anything-as-a-Service (XaaS): a further method of product that can be something or whatever as a product is the XaaS. The database system will be able to use Security-as-a-Service, ethnicity-as-a-Service, Communications-as-a-Service, DaaS (Database-as-a-Service) or tactic-as-a-Service and so on to manage the enormous quantity of funds to meet intimate, granular and precise requirements [62].
Fault resistance models in disseminatedorganizations
For a device, condition monitoring is essential in order to allow it to provide the required services even in the face of system failure and one or more faults [30,31]. Errors in a process emerge as a result of medical errors that are attributable to errors in turn. This is defined as follows:
Faults: A system's failure to perform a required function / necessary task is triggered by some irregular state or error occurring in one or more components of a system.
Error: Because of the existence of faults, a device component may switch to an error message or an inaccurate condition. The erroneous output of a storage system could lead in a significant or even total failure of the system. Failure: refers to the platform's misbehavior that even a user (a person or any another computer chip) may detect. A malfunction is documented only if the performance or consequence of the system that's wrong. Strategies to remitted are important because they help to identify and manage device faults that can occur whether attributable to hardware (H / W) malfunction or software (S / W) faults [32]. In the cloud framework, condition monitoring is particularly
important as it provides confirmation of performance, reliability, and enable access. In order to attain the robustness of cloud services, the malfunction must be easily accessed and controlled.
Reactive high availability: After the weaknesses / flaws have already happened, this technique is primarily used to reduce the impact of error in the cloud service. It gives a scheme effectiveness or consistency. For the cloud and other communications databases, proactive fault tolerance techniques were investigated.
Hardware fault-tolerance:-This technique is being used to proactively anticipate the failures and replace other running elements for the suspect component, i.e. it prevents fault and failure repair.
Error resistance in disseminated computing atmospheres
Fault tolerance (FT) is an important problem in the cloud platform because it helps the framework to even provide great result for the necessary services throughout the existence of one or even more storage system failures [33,34]. Other than cloud services, fault - tolerant techniques have been suggested to several various distributed computing environments in the background. A few of the following are just as continues to follow:-
Wired large geographical:-It is a series of autonomous computers that behave to its customers / users as a single coherent system. An independent collection of resources is included in all machines in the distributed network and certain general network adapters may be shared, such as a printer. For connectivity in a distributed network, message forwarding is commonly used. Attributed to the prevalence of modules that may have been installed at various locations / sites, developing a distributed network is a challenging job. FT is among the big obstacles that the creator of the device has to face. In particular, in decentralized communication networks, FT is strongly needed, particularly in a large-scale setting. Dispersed device users need the scheme to maintain progressively proficient, even in the event of technological malfunctions [35]. When one or more of the program's members have failed, the team shall be capable to face the necessities of a client then again. Consequently, to accommodate the inadequate product of its elements, an effective framework should be designed and managed.
The most basic methods for FT in embedded environments are malfunction detection (FD) and quality control. To deal with failures in all of these systems, reactive fault - tolerant methods like checkpoint, duplication, retry, resubmitting, etc were used.
Mobile CalculatingModel: - This is a different network sort, whereby desktop devices is among or all of constituent’s points. Even in the presence of host' flexibility, this device retains constant internet connectivity since their position inside the system can differ over time. With intermittent asynchronous data contact, each branch in the framework works independently. A fixed channel may be used to connect up the static points in the mobile device. In addition, to create connectivity here between connecting nodes, i.e. the remote node as well as the other nodes within the device, a static node (frequently the mobile core network) has been used. Throughout the mobile device, users transmit using communications with one another.
Many mobile device limitations include restricted capacity, restricted disc space for network devices, mobility support, short battery life, etc. Many solutions to fault sensitivity are being utilized to address the limitations of the personal computing framework [36]. Verification-pointing is the most widely utilized FT solution in the specific phone, as available supplies prohibit the use of complication-based systems such as duplication, etc. To transfer the error-free state to any processing (fixed / constant), the FT technology involves procedures which are try-pointed at frequent intervals. If there is some error in the method, by locating the latest saved / preserved state (named rollback turnaround), the loss can be retrieved.
It is possible to identify higher in terms schemes as organised, interaction-induced, and disorganised checkpointing [37]. In a synchronised way, by transmitting the security check-based coordination signals, the structures adapt their checkpointing behaviour. The organised check-pointing strategies have huge footprint notifications, so they are not ideal for digital devices but have very poor digital modems channels / channels in the bandwidth utilization. In addition, the implementation of the procedure will also need to be halted only at time of bottleneck management that may results in output deterioration. The uncoordinated procedures of checkpointing allow procedures to obtain checkpoints lacking coordination with those at frequent intervals, however this system can suffering from of the chain reaction [38, 39]. To tackle the chain reaction, an interaction-induced test-pointing technique was used.
Phone-Grid Computing:-Grids, decentralized in existence, are quite large-scale structures. The volume of remains to be undertaken amongst these protocol stacks was distributed by these. Grid computing encourages the integration of loosely affiliated, embedded environments with large-scale support to satisfy the computing requirements of large-scale activities. Grid computing thus offers tremendous computation, bandwidth capacity and processing for consumers. To get better performance, it also is necessary to use computing systems in combination with mobile technology. This strategy is also critical in order to effectively deal with vital mobile device restrictions [40]. Nevertheless, because of poor links, random node flexibility, battery reliance, small capacity for communication, limited ability for transmission and fixed space, the integration of mobile and network devices for
the use of computational power is difficult. Actual management of mobile grid computing (MoG) software systems is beneficial if the flaws / failures of smart phones are treated correctly [41]. In order to manage the various faults throughout the MoGs, FT regulations also are necessary. In MoGs, the most widely used FT tool is checkpointing and rollback retrieval. In traditional wired and wireless modular decentralized networks, these FT approaches were used thoroughly.
MANET (Mobile ad hoc networks(N/W)): - This is a self-configuring wireless sensor network and does not rely on networks , i.e. it is a networks-less wirelessly linked mobile phone network. All instruments in the MANET are independent and can quickly adjust their course and direction, thus constantly altering connections to many other machines. MANETs are commonly utilised to increase current mobile networks and MoG (mobile grid computing) computational power. However, some transient / momentary and irreversible failures are susceptible to MANETs. An FT tool is necessary for use successfully to manage the failure [42]. In static and/or cellular mobile networks, checkpointing, and also rollback restoration, is a commonly utilized policy to tackle defects. The usage of FT methods with MANETs, nevertheless, is less discussed. The following recovery-based strategies really aren't directly applied to the MANETs due to certain difficulties like static connectivity deficit, frequent node motion, limited bandwidth, and limited reliable storage [43,44]. Rollback retrieval typically encompasses on checkpointing, which specifies the intensity of the checkpoints of a smart phone / node dependent on the movement / portability, has also been created to optimize defects in MANETs and thus prevents excessive checkpoints [45].
Errorresistance methods in the cloud computing
Via the Network, computing offers different services and flexible computing power. A DC (data centre) on the supplier's side offers the capability to manage computer networks but also their relevant parts, such as networking , storage, uninterruptible power, etc. Many virtual machines (VMs) operate on actual machines in order supply management consulting.
DC Cloud [46]. Various kinds of communication networks are used by these DCs. In cloud computing, fault - tolerant methods rely on embedded communication networks. The network topology is the configuration of points within the same network. In another words, configuration is a fundamental network key component that links computer systems to each other [47,48]. Bus, ring, star, grid, tree, and hybrid topology are also the simple technologies of the system.
Bus Topology:-This architecture is used for the patch cord link of all computers and communication devices. In one step, it transmits information. This topography, used during local applications, is quite cost-effective, simpler to grasp and extend. Nevertheless, when cable break, the whole network fails[49] throughout this geometry. The efficiency of such a configuration is reduced if the channel is strong. A tiny quantity of length is included in the cable. Slows down relative to this topology, according to topology with loop.
Ring Topology: - Computer devices are linked to one another in a covalent bond in this topology, how the last unit is linked to the other one. This architecture can be built and also expanded very cheaply. The transmission system is unchanged in the event of heavy internet traffic as well as the addition of any additional nodes [50]. The malfunction of one computer network, nevertheless, can influence the entire network. Through adding or removing the computers, the network activities are interrupted. In wired networks, problem solving is quite difficult as well. Star Topology:-This circuit has been used with any aid of a cables to link all computers to a centralized point. As a coordinator hub, such hub is used and all other networks accessible are connected to it. This topology offers fast efficiency, simpler to figure out how to solve, configure and modify [51]. These topology, nevertheless, was costly to utilize so if the hub fails, so anwhole system fails to work. The cost of maintenance is large.
Mesh Topology: - All of the nodes or computer networks are completely connected to one another in these configuration. This topography is rather stable, which makes the malfunction simpler to diagnosis. Privacy and confidentiality are also supported [52]. Nevertheless, installation or configuration of this configuration is quite challenging. Often, the price of power cables is high and mass wiring is necessary.
Tree Topology:-An input space is found in this configuration and other such processors or nodes (recognized as hierarchical configuration) are connected to it. The hierarchy's average amount must be three. This configuration is an expanded variant of the topology for buses and stars [53]. Managing and managing it is often simpler. Error correction is also readily performed in this configuration. This topology, however, involves a topology that it is an expensive operation and extensively cabled.
Hybrid Topology: -These network is a amalgamation of two technologies or more than two. Features such as reliability, scalability, and versatility are given by this geometry. Its architecture, nevertheless, is complex involving a costly operation.
Conclusion
The cloud model has now become a fact and academics, the IT industry as well as other organisations have embraced it. Methods to remitted are needed to enhance quality of the service in the cloud world. Nevertheless, some simple metrics, like reaction time, accessibility, throughput and efficiency, have only been used to answer the reliability problem of simple database business processes and assessment. Currently, in both industry and academia, the cloud computing paradigm is now the most appealing computing platform. Unique variations such as scalability, permeability, highly available and many others are supported by cloud services. The IT industry has been transformed by the cloud computing paradigm as it offers many advantages to people, academics, organisations, and even countries. The cloud infrastructure is also vulnerable to errors, despite having multiple benefits. In cloud computing, errors are unavoidable relation to the large of service. Fault detection policies and regulations introduced in the private cloud to manage faults efficiently. Technologies of fault - tolerant help to avoid and accept device faults, that can occur either attributable to hardware or software failure. The primary motive for using cloud computing fault - tolerant strategies is to accomplish failure restoration, highly reliable, and improve availability.
References
1. P. Hansen and N. Mladenovic, “Variable neighborhood search: Principlesand applications,” European Journal of Operational Research,vol. 130, no. 3, pp. 449-467, 2001.
2. C. Min and Y. Eom. “Integrating Lock-Free and Combining Techniquesfor a Practical and Scalable FIFO Queue,” IEEE Transactionson Parallel and Distributed Systems, vol. 26, no. 7, pp. 1910-1922, 2015. 3. G. D. Pietro and A. Coronato, “Tools for the Rapid Prototyping ofProvably Correct Ambient Intelligence
Applications,” IEEE Transactionson Software Engineering, vol. 38, no. 4,pp. 975-991, 2012.
4. R. Ssembatya and A. V. D. M. Kayem, “Secure and Efficient MobilePersonal Health Data Sharing in Resource Constrained Environments,”IEEE International Conference on Advanced Information Networkingand Applications Workshops (WAINA), pp. 411-416, 2015.
5. J.Winkley, P. Jiang, andW. Jiang, “Verity: an ambient assisted livingplatform,” IEEE Transactions on Consumer Electronics, vol. 58, no. 2,pp. 364-373, 2012.
6. T. Magherini, A. Fantechi, C. D. Nugent, and E. Vicario, “Usingtemporal logic and model checking in automated recognition ofhuman activities for ambient-assisted living,” IEEE Transactions onHuman-Machine Systems, vol. 43, no.6, pp. 509-521, 2013.
7. J. C. Wang, C. H. Lin, E. Siahaan, B. Chen, and H. Chuang, “Mixedsound event verification on wireless sensor network for homeautomation,” IEEE Transactions on Industrial Informatics, vol. 10, no.1, pp. 803-812, 2014.
8. B. Y. Xu, L. D. Xu, H. M. Cai, C. Xie, J. Hu, and F. Bu, “Ubiquitiousdata accessing method in IoT-based information system for emergencymedical services,” IEEE Transactions on Industrial Informatics,vol. 10, no. 2, pp. 1578-1586, 2014.
9. C. H. Liu, J. Wen, Q. Yu, B. Yang, and W. Wang, “HealthKiosk: Afamily-based connected healthcare system for long-term monitoring,”IEEE INFOCOM, pp. 241-246, 2011.
10. M. Barua, X. Liang, R. Lu, and X. Shen, “PEACE: An efficientand secure patient-centric access control scheme for eHealth caresystem,” IEEE INFOCOM, pp. 879-886, 2010.
11. A. Bamis, D. Lymberopoulos, T. Teixeira, and A. Savvides, “TheBehaviorScope framework for enabling ambient assisted living,”Personal and Ubiquitous Computing, vol. 14, no. 6, pp. 473-487, 2010.
12. M. Ruta, F. Scioscia, G. Loseto, and E. D. Sciascio, “Semantic-basedresource discovery and orchestration in Home and Building Automation:a multi-agent approach,” IEEE Transactions on IndustrialInformatics, vol. 10, no. 1, pp. 730-741, 2014.
13. H. Mei, B. J. Beijnum, I. Widya, V. Jones, and H. Hermens,“Enhancing the performance of mobile healthcare systems basedon task-redistribution,” IEEE INFOCOM, pp. 1-6, 2008.
14. N. K. Suryadevara and S. C. Mukhopadhyay, “Determination ofWellness of an Elderly in an Ambient Assisted Living Environment,”IEEE Intelligent Systems, vol. 29, no. 3, pp. 30-37, 2014.
15. M. A. Hossain, P. K. Atrey, and A. E. Saddik, “Modeling andassessing quality of information in multisensor multimedia monitoringsystems,” ACM Transactions on Multimedia Computing, Communications,and Applications, vol. 7, no. 1, pp. 33-63 2011.
16. L. Cheng, J. Niu, Y. Gu, C. Luo, and T. He, “Achieving EfficientReliable Flooding in Low-Duty-Cycle Wireless Sensor Networks,”IEEE/ACM Transactions on Networking, vol. 24, no. 6, pp. 3676-3689,2016.
17. F. Z. Benhamida, and Y. Challal, “FaT2D: Fault Tolerant DirectedDiffusion for Wireless Sensor Networks,” IEEE International Conferenceon Availability, Reliability and Security, pp. 112-118, 2010. 18. S. K. Singh, M. P. Singh, and D. K. Singh, “A survey of energyefficienthierarchical cluster-based routing
in wireless sensor networks,”International Journal of Advanced Networking and Application,vol. 2, no. 2, pp. 570-580, 2010.
19. I. Stojmenovic and W. Sheng, “The fog computing paradigm: Scenariosand security issues,” IEEE Federated Conference on ComputerScience and Information Systems, pp. 1-8, 2014.
20. Q. Hu, C. Wu, X. Zhao, X. Chen, Y. Ji, and T. Yoshinaga, “VehicularMulti-access Edge Computing with licensed Sub-6 GHz, IEEE802.11p and mmWave,” IEEE Access, Vol. 6, no. 1, pp. 1995-2004,2017. 21. Y. Cao, S. Chen, P. Hou and D. Brown, “FAST: A fog computingassisted distributed analytics system to
monitor fall for strokemitigation,” IEEE International Conference on Networking, Architectureand Storage, pp. 2-11, 2015.
22. M. Aazam and E. Huh, “E-HAMC: Leveraging Fog computing foremergency alert service,” IEEE International Conference on PervasiveComputing and Communication Workshops, pp. 518-523, 2015. 23. M. S. Mahmud, U. Sarker, M. M. Islam, and H. Sarwar, “A GreedyApproach in Path Selection for DFS
Based Maze-map DiscoveryAlgorithm for an autonomous robot,” IEEE International Conferenceon Computer and Information Technology, pp. 546-550, 2012.
24. F. Busato and N. Bombieri, “BFS-4K: an Efficient Implementationof BFS for Kepler GPU Architectures,” IEEE Transactions on Paralleland Distributed Systems, vol. 26, no. 7, pp. 1826-1838, 2015.
25. Liu, H., Xu, D., Miao, H., 2011. Ant colony optimization based service flow schedulingwith various QoS requirements in cloud computing. In: First ACIS InternationalSymposium on Software and Network Engineering. IEEE, pp. 53–58.
26. Macskassy, S.A., Provost, F., 2007. A brief survey of machine learning methods forclassification in networked data and an application to suspicion scoring. In:Statistical network analysis: models, issues, and new directions. Springer,Berlin, Heidelberg, pp. 172–175.
27. Mell, P., Grance, T., 2011. The NIST definition of cloud computing recommendationsof the national institute of standards and technology. Nist Spec. Publ. 145, 7.
28. Mohanapriya, N., Priya, C., 2016. A fault tolerant router with optimized an colonyalgorithm. In: Advanced Communication Control and Computing Technologies(ICACCCT), 2016 International Conference on. IEEE, pp. 167–170.
29. Mukwevho, M.A., Celik, T., 2018. Toward a smart cloud: a review of fault-tolerancemethods in cloud systems. IEEE Trans. Serv. Comput. 1374 (c), 1–18.
30. Nazari Cheraghlou, M., Khadem-Zadeh, A., Haghparast, M., 2016. A survey of faulttolerance architecture in cloud computing. J. Netw. Comput. Appl. 61, 81–92.
31. Pandey, S., Wu, L., Guru, S.M., Buyya, R., 2010. A particle swarm optimization-basedheuristic for scheduling workflow applications in cloud computingenvironments. In: 24th IEEE International Conference on AdvancedInformation Networking and Applications, pp. 400–407.
32. Pandya, K., 2013. Network structure or topology. Netw. Struct. Topol. 1 (2), 22–27.
33. Park, J.H., Park, J.H., 2017. Blockchain security in cloud computing: use cases,challenges, and solutions. Symmetry 9 (8), 164.
34. Park, T., Woo, N., Yeom, H.Y., 2002. An efficient optimistic message logging scheme forrecoverable mobile computing systems. IEEE Trans. Mob. Comput. 1 (4). 265–251.
35. Park, J., Yoo, G., Lee, E., 2005. Proactive self-healing system based on multi-agenttechnologies. In: Software Engineering Research, Management andApplications, 2005. Third ACIS International Conference on. IEEE, pp. 256–263.
36. Patidar, S., Rane, D., Jain, P., 2011. A survey paper on cloud computing. In: Proc. -2012 2nd Int. Conf. Adv. Comput. Commun. Technol. ACCT 2012, pp. 394–398.
37. Patil, A., Shah, A., Gaikwad, S., Mishra, A.a., Kohli, S.S., Dhage, S., 2011. FaultTolerance in Cluster Computing System. In: 2011 Int. Conf. P2P, Parallel, Grid,Cloud Internet Comput., pp. 408–412.
38. Poola, D., Ramamohanarao, K., Buyya, R., 2014. Fault-tolerant workflow schedulingusing spot instances on clouds. Procedia Comput. Sci. 29, 523–533.
39. Prakash, R., Member, S., Society, I.C., 1996. Low-cost checkpointing and failurerecovery in mobile computing systems. IEEE Trans. Parallel Distrib. Syst.(October), 1035–1048.
40. Prathiba, S., Sowvarnica, S., 2017. Survey of failures and fault tolerance in cloud. In:Proc. 2017 2nd Int. Conf. Comput. Commun. Technol. ICCCT 2017, pp. 169–172.
41. Qiu, F., Zhang, B., Guo, J., 2016. A deep learning approach for VM workloadprediction in the cloud. In: 2016 17th IEEE/ACIS International Conference onSoftware Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD). IEEE, pp. 319–324.
42. Qiu, W., Zheng, Z., Wang, X., Yang, X., Lyu, M.R., 2014. Reliability-based designoptimization for cloud migration. IEEE Trans. Serv. Comput. 7 (2), 223–236.
43. Rani, A., 2016. Proposed model for multi-objective differential evolution (mode) incloud computing. Int. J. Adv. Res. Innovative Ideas Educ. 3, 4119–4124.
44. Rimal, B.P., Choi, E., Lumb, I., 2009. A taxonomy and survey of cloud computingsystems. In: NCM 2009 - 5th Int. Jt. Conf. INC, IMS, IDC, pp. 44–51.
45. Saikia, L.P., Devi, Y.L., 2014. Fault tolerance techniques and algorithms in cloudsystem. Int. J. Comput. Sci. Commun. Networks 4 (1), 1–8.
46. Santra, S., Acharjya, P.P., 2013. A study and analysis on computer network topologyfor data communication. Int. J. Emerg. Technol. Adv. Eng. 3 (1), 522–525.
47. Savu, L., 2011. Cloud computing: deployment models, delivery models, risks andresearch challenges. In: 2011 Int. Conf. Comput. Manag. CAMAN 2011.
48. Sem-Jacobsen, F.O., Skeie, T., Lysne, O., Duato, J., 2011. Dynamic fault tolerance in fattrees. IEEE Trans. Comput. 60 (4), 508–525.
49. Sharma, S., 2017. Enhance data security in cloud computing using machine learningand hybrid cryptography techniques. Int. J. Adv. Res. Comput. Sci. 8 (9), 393–397.
50. Shaw, R., Howley, E., Barrett, E., 2017. An Advanced Reinforcement LearningApproach for Energy-Aware Virtual Machine Consolidation in Cloud DataCenters, pp. 61–66.
51. Shwe, T., Aye, W., 2008. A fault tolerant approach in cluster computing system. In:5th Int. Conf. Electr. Eng. Comput. Telecommun. Inf. Technol. ECTI-CON 2008,vol. 1, pp. 149–152.
52. Singh, P., Cabillic, G., 2003. A Checkpointing Algorithm for Mobile ComputingEnvironment, 65–74. 53. Singh, S., Jeong, Y.S., Park, J.H., 2016. A survey on cloud computing security: issues,threats, and
solutions. J. Netw. Comput. Appl. 75, 200–222.
54. Singh, G., Kinger, S., 2013. A survey on fault tolerance techniques and methods incloud computing. Int. J. Eng. Res. Technol. 2 (6), 1215–1217.
55. Sridaran, R., 2015. An overview of DDoS attacks in cloud environment. Int. J. Adv.Networking Appl. (IJANA), 124–127.
56. Strayer, W.T., Lapsely, D., Walsh, R., Livadas, C., 2008. Botnet detection based onnetwork behavior. In: Lee, W., Wang, C., Dagon, D. (Eds.), Botnet Detection.Advances in Information Security, vol 36. Springer, Boston, MA.
57. Tantikul, T., Manivannan, D., 2005. A communication-induced checkpointing andasynchronous recovery protocol for mobile computing systems. Parallel Distrib.Comput. Appl. Technol. PDCAT Proc. 2005, 70– 74.
58. Tosh, D.K., Shetty, S., Liang, X., Kamhoua, C.A., Kwiat, K.A., Njilla, L., 2017. Securityimplications of blockchain cloud with analysis of block withholding attack. In:Proceedings of the 17th IEEE/ACM International Symposium on Cluster, Cloudand Grid Computing. IEEE Press, pp. 458–467.
59. Trivedi, M., 2016. A survey on resource provisioning using machine learning incloud computing. Int. J. Eng. Dev. Res. 4 (4), 546–549.
60. Tsai, J.T., Fang, J.C., Chou, J.H., 2013. Optimized task scheduling and resourceallocation on cloud computing environment using improved differentialevolution algorithm. Comput. Oper. Res. 40 (12), 3045–3055.
61. Umamakeswari, P.V.M.D.A., Shaileshdheep, S.G.B.G., Aishwarya, V., Subramanian, E.R.S., 2017. A study of various meta heuristic algorithms for scheduling in cloud.Int. J. Pure Appl. Math. 115 (7), 205– 208.
62. Valle, G. et al., 2008. A framework for proactive fault tolerance. In: ARES 2008 - 3rdInt. Conf. Availability, Secur. Reliab. Proc., pp. 659–664.
