14
Convergence of Big Data and Cloud Computing Environment

Ranjan Ganguli

All India Institute of Local Self Government, Deoghar, India

Abstract

Big Data is created every day from various heterogeneous sources over the internet and is a challenging task in processing, handling and storing for major organization. To be dealt with, convergence of Cloud computing with Big data has played a vital role as a solution of cost effective, better flexibility with data security at a higher level. This creates an edge over traditional computing methods and delivered at a cheaper rate. This Chapter introduces Big data and its challenges in respect of traditional databases. In various other parts, we will discuss how convergence activities played a major role with Big data in finding a solution in Cloud computing platform for handling large volume, fast moving and ever increasing data in the internet along with its pros and cons of the system. Finally, future aspects of Big data with Cloud computing will be discussed.

Keywords: Big data, Cloud computing, Hadoop, Cassandra

14.1 Introduction

The term big data is common to IT industry or business companies nowadays which collects data from hundreds of heterogeneous sources, including social media, email correspondence, video posts or tweets and credit card data. So, data is becoming huge and need special attention to analyze and analytic engine for greater efficiency and at reduced cost. Organizations need to have special value chain system to coordinate effective activities. This can only be achieved through integrated systems that can manage smooth data flow from devices to data center [1]. In various small to large scale enterprises, organizations are moving rapidly towards cloud-based solutions for taking advantages of scalability, QoS (Quality-of-Service) management, privacy and security. This type of closed services in the market reduces infrastructure and operational cost.

Schematic illustration of the next generation convergence technologies.

Figure 14.1 Next generation convergence technologies.

Like SQL-based relational data model generated a platform for expressing various requirements options and cooperate customers to choose from several vendors and in converse, it increases the competition rate of selection process. The main problem is to handle and integrate the present challenges faced by the existing feature and in big data area. As was discussed in previous stage of cloud-based services, network technology was considered as a separate thing with data center and the associated client. But Lamport’s [2] theory on distributed computing are widely accepted and recognized as a vital step for problem solving of consistency and persistency with fault-tolerance mechanism. Thus an integrated platform shown in Figure 14.1 must be available to increase the potential of investment with decrease in overall application development cost for more effective data handling and management from device to the data center directly. Several organizations are busy creating a model of leveraging several application development processes in respect of mobility awareness solutions to deliver and create sophisticated end-to-end secure user experiences. Only specific group members can access the private clouds that exist for a certain group of organizations.

Cloud technologies use different APIs to process Big Data in a simpler way, just like Amazon EC2 virtual machine’s capacity is 1 to 20 processors, with up to 15 GB of memory and 160 TB of storage and uses Web.

Service (WS) portal. Presently, Cloud-based toolkits manage virtual machines directly because of per flexible requirement of consumers. So, to look ahead, a need of re-structuring of internet service environment is required with people preferring more personalized services.

14.2 Big Data: Historical View

The general historical view of big data started 5,000 years back to Mesopotamia, when they keep track of their accounting information and business activities like crop growth and herding in a specified manner. The principles continued to grow until in 1663, when John Graunt, was able to record and analyze the mortality rate of the bubonic plague that happened during that time in London. He did this effort to raise a concern and awareness among the people. This was the first statistical analysis of data ever recorded in the world. With this revolution of taking care of data until 1889 (also starting point of modern era), Herman Hollerth made an attempt to organize census data using a computing system. After his great effort, the various other next noteworthy data development continued until 1937, when US government contracted with IBM (International Business Machine) to develop a punch card reading system to keep track of millions of Americans. Also, the first data processing machine was developed in 1943 and named ‘Colossus’ was used to decipher Nazi codes and to identify or search for patterns to appear in intercepting regular messages during Second World War. But with the development of first data center in US in 1965, which had stored millions of tax returns and fingerprint sets has made a significant impact in conversion of paradigm from data to information, when first introduced the term ‘Big Data’ by Roger Mougalas in 2005. In 2010, Eric Schmidt, executive chairman of Google, told a conference in Lake Tahoe that the volume of data is being generated every two days is same as was created from the beginning of time up to the year 2003 and created a new leap of information era. Later on in 2012, the definition of big data included 3Vs as Volume, Velocity and Variety to understand the characteristics of big data. There are many other interesting changes that will appear in Big Data revolution in the next few years. International Data Corporation (IDC) which defined big data as “a new generation of technology and architecture that is designed to collect or extract meaning or value from a large volume set of wide variety of data, by enabling the high velocity capture, discovery or analysis,” pointed out that the 3Vs of big data can be extended to 4Vs by adding ‘Value’ namely as ‘Volume’, ‘Variety’, ‘Velocity’ & ‘Value’ and creates a new definition of big data that is widely recognized because of its highlighting the necessity of data.

14.2.1 Big Data: Definition

Based on several observations and various aspects in respect of big data, it can be generally defined as “a new form of data integration to uncover hidden patterns from large datasets that are diverse (structured, semi & unstructured), complex and massive in scale” as shown in Figures 14.2 and 14.3.

  1. (i) Volume: It refers to varieties (heterogeneous form) and different sizes of data that is continuously generated from various sources for finding patterns from information through analysis. Nokia motivated, mobile data challenge initiative collects longitudinal data (that requires efforts and investments) from various smart phones and made this available for research in community [3]. This challenge produced an interesting result similar to the predictability of human behaviour patterns for complex data.
    Schematic illustration of the comparison of four versus big data in data scientist view.

    Figure 14.2 Four vs of Big data: Data scientist view.

    Schematic illustration of the classification of big data.

    Figure 14.3 Big data classification.

  2. (ii) Variety: When data is collected via various sources like phones, sensors, social networks-data appears in various heterogeneous format or data types like structured, semi structured and unstructured. Like, most of the mobile data and sensors collected data is unstructured in format along with internet users that generate randomly, all set of structured, semi & unstructured data [4].
  3. (iii) Velocity: It describes the speed of incoming and outgoing data with the time taken to process and analyze as streamed data are collected across time in multiple sources, legacy collections, predefined achieved data.
  4. (iv) Veracity: It is not only the measure of the quality of data but also the trustworthiness (types, source and analysis) of data.

14.2.2 Big Data Classification

For better understanding of characteristics, Big data is classified in different categories: i) Sources, ii) Structure format, iii) Stores, iv) Data Building, and iv) Data Handling.

The complexities of each category have been mentioned in Figure 14.3. Various data sources like internet data, sensing data, transactional data and others are shifted from unstructured to structured form and stored in various formats like the relational database which comes in many varieties.

14.2.3 Big Data Analytics

We know that big data deals with data set of higher dimension in terms of volume, diverse contents including structured, semi-structured and unstructured data (as variety), arriving and processing at a faster rate (as velocity) than any small to large scale organization had to deal or face with such as oceanic volume of data produced through several devices connected across any network like PCs, smart phones to sensors including traffic cameras that arrive under various formats. The significance of big data comes when this is analyzed—discovered or finding hidden patterns, helping decision makers with broader sense of adaptability and ability to handle and respond to the world in a more challenging situation with uses of more intelligence-based algorithm or through recent approaches like machine learning, artificial neural network, computational mathematics, etc. to move through data for discovering interrelationship among data. So, for Analytics:

  1. (i) Data is more important than ever—like what can be done with the data? – medium to large enterprises looking to unlock data’s competitive advantage
  2. (ii) Moving Data analytics from batch to real time systems
  3. (iii) Predictive analytics under real time basis
  4. (iv) Scope of big data analytics increases.

Thus, data processing and analytics can be applied to cloud-based solutions, IT support infrastructure with high end computation and capacity solutions.

14.3 Big Data Challenges

  1. (i) Data Collection

    The important thing about big data is its increasing data collection phase with increase in velocity all the time and brings sometimes a major issue in delivering the data into cloud for further processing. Existing internet standard would create a bottleneck for cloud to perform services in the cloud and need next level of supporting standard or techniques to deal with present scenario for higher efficiency and smooth data traversal across or to the cloud system.

  2. (ii) Storage

    Today’s legacy database systems are not able to find a way to grab possible advantages of scalability, present in cloud’s environment for optimal state or use. Thus, a scalable model is required in new systems under consideration.

  3. (iii) Analysis

    Data analysis helps to find value from big data. Modern Techniques and methods must be researched and available to process rising data sets. Also, simplification must be a goal of consideration during the analysis phase of big data.

  4. (iv) Security

    Two main challenges need to focus in securing big data in large systems: a) Limited expenses to be carried out without effecting the performance. b) Extensive experimentation or testing is required on data analysis process to identify possible attacks on mostly used RDBMS, NoSQL.

    Schematic illustration of a big data view on smart city formation.

    Figure 14.4 A big data view on smart city formation.

  5. (v) Lack of skilled professionals

    Big data analytics requires extensive knowledge about data and portion of statistics to understand the problems and providing a better business decision. Skilled data scientist and analyst with in-depth knowledge about analytics tools are less in the market and the demand will grow with adoption of large scale integration (Figure 14.4).

In 2015, a survey conducted to address the problems faced in big data working environment is shown in Figure 14.5.

14.4 The Architecture

The Big data architecture is a multilayer processing system which consists of various components that includes data sources & storage, batch processing, real-time messaging sources, data streaming, analytical storage with reporting and orchestration. But the abstract view of the whole thing can be categorized as three independent layers as i) storing—that generally holds the data, ii) Handling—done as the mid-level layer, and iii) Processing—at the upper level that support various tools that meet the end user. Below Figure 14.6 shows the convergence abstract view.

The platform used is No SQL or Parallel DBMS. Map Reduce supports a parallel processing of large volume of data after dividing the user query and sent it to different nodes in cloud. A prioritized ranking system of result with reduced phase is done and integration is maintained using DBMS technology.

Bar chart depicts programmers challenges with big data analytics.

Figure 14.5 Programmers challenges with big data analytics.

Schematic illustration of cloud convergence with Big data in its architecture.

Figure 14.6 Cloud convergence with Big data in its architecture.

14.4.1 Storage or Collection System

There is a remarkable improvement in capacity technologies in dealing with Big data to deal with data rehydration. The meta data management requires high end storing capacity and bandwidth. Sometimes, a single centrally handled file is responsible to deal with millions of files and meta data records. Distributed storage share between processors in the system is a big challenge for a storage management. So, data that can be stored and measured either it is customer information, service log or call record generated by telecom industries, sensor data, user created data (social data), blogs, various mails, etc., everything. Multicore processors have additional cycles to combine that meet with storage systems. Presently, cloud based storage systems that supports distributed file division exist like HDFS (Hadoop Distributed File Systems), Amazon S3 or OpenStack Swift and Google File System (GFS or GoogleFS) (Table 14.1) for storing huge data in network along with storage area network and network attached storage.

14.4.2 Data Care

With latest Map and Reduce technology to handle and processed in parallel a large set of data in a cluster basis, it acts as a framework in Apache Hadoop—that contributes sched-uling with distribution and configuration that constitute the two-step process of map and reduce. Different computers are tied up with a weak correlation mechanism and expanded up to thousands of different computers. So, with no exception, a possibility of system errors is in common. Thus, Map and Reduce-based simplified operations are helpful in solving complicated problems without knowing programmer a deep understanding of parallel programming of handling. Several computers connected in the system supports maximum throughput. Finally, data that is stored in HDFS and is divided to worker and that can be further expressed to a value type with result collected in local hardware (disk). It is then assembled and produced in a separate file.

14.4.3 Analysis

KDD (Knowledge Discovery in Databases) is the process of finding meaningful information in data. It is actually to store the data, process it and analyze it for discovering hidden facts that are unknown so far and converted in knowledge for decision making.

Table 14.1 Distributed file system: On cloud services.

Examples Descriptions
Google FS (GFS) A scalable distributed or give out file system that use cluster based commodity hardware for huge data intensive applications with fault handling mechanism.
Hadoop File System An open source distributed system for huge data computation and operated under MapReduce model with low cost usage.
Amazon Simple Storage Service (S3) An online public storage web services offered by Amazon.

14.5 Cloud Computing: History in a Nutshell

The timeline of cloud computing showed that the history began in early 60s when computers were considered to be used by more than two people and tried to fill up empty space between user and various service provider. But, later in early 2000s when Amazon introduced their web-based retail services that incorporated cloud computing model and enhanced the computer capacity with more efficiently. The term “Cloud Computing” in the conference was described as—a modern trend where people can access the power of software and hardware applications over any web in replace of their personal desktop computers [5]. Soon after that, large organizations like Google, IBM have started developing cloud supported services and by 2012, Oracle introduced ‘Oracle Cloud’ with a well-defined architecture that offered three basics of business as IaaS (Infrastructure-as-a-Service), PaaS (Platform-as-a-Service) and SaaS (Software-as-a-Service).

14.5.1 View on Cloud Computing and Big Data

National Institute of Standards and Technology (NIST) defines cloud in respect of its characteristics as “Cloud computing is a model for convenient on shared pool of configurable computing resources (like: networks, servers, storage, services, etc.) that released services with minimal management effort.”

Cloud models help organizations to evaluate best business strategy and user requirements. Like an organization can use or add analytics to their own private cloud environment to protect sensitive information with extends to hybrid cloud systems and can take benefits of other data sources come up in public clouds. As per IDC prediction, by the year 2020, 40% of the world data will be in cloud as big cloud-operated companies like Amazon, IBM and Microsoft have already deployed their big data based activities using Hadoop clusters. The relationship between big data and cloud based operations [6] is shown in Figure 14.7.

The Cloud comes up with analytics a high capacity and more processing power with use of large datasets to return more useful information.

14.6 Insight of Big Data and Cloud Computing

User needs can be addressed with full range of data analytics on data that includes Analytics-as-a-Service (AaaS) that support business intelligence to go through an alternative way to deal with software and internal hardware solutions with wide range in delivery to usage of data. In Enterprise-based solutions, data can be optimized with AaaS supporting capabilities as:

  • – Extract semi, structured and unstructured forms of data from several sources.
  • – Maintaining all data supported guidelines and policies for control activities.
  • – Performing activities to integrate data, performing analysis, transformation with visualization.
Schematic illustration of the application of Big data in cloud services.

Figure 14.7 Application of Big data in cloud services.

14.6.1 Cloud-Based Services

The basic cloud-based services [7] as described in Figures 14.8 and 14.9 are:

  1. a) Infrastructure-as-a-service or IaaS
  2. b) Platform-as-a-Service or PaaS
  3. c) Software-as-a-Service or SaaS.
  1. a) Deployment of IaaS (also called Hardware-as-a-Service) is done through a cloud provider that helps to allocate time shared server resources that normally fulfils the required performance or computational and storage needs for all kinds of analytical purposes. Basically, cloud-based operating systems take care of all kinds of networks performance and storage issues. It is a foundation for many companies’ cloud-related services with strong involvement of investment and IT support services to implement big data analytics. Organizations install their own software in their platform either like Hadoop framework, or NoSQL database, Apache Cassandra, etc. Resources are managed with automated tools and can be made easier for resource orchestration. List of IaaS service providers in the market technology are:
    Schematic illustration of utility model of cloud based computing.

    Figure 14.8 Utility model of cloud based computing.

    Schematic illustration of three layered of cloud as a service model.

    Figure 14.9 Three layered of cloud as a service model.

    • Web Services-Amazon
    • OpenStack Software
    • VMware vCloud Suite
    • Windows Azure.
  2. b) Tools and libraries are available for developers for deployment, testing, building and run several applications on cloud supported infrastructure. This puts down various management workloads that Hadoop requires to scale and configure process and served as a development platform for various analytics applications. List of PaaS supported services in cloud related technologies are:
    • Open Shift-Red Hat
    • Azure-Windows
    • Google Compute Engine (GCE)
    • Magento Commerce Cloud.

    The development environment of PaaS is not hosted locally and opens a window for developers to access it from anywhere in the world. It can be accessed over any internet platform and can create an application in a web browser. So, there are no geographic restrictions for any team to collaborate in the process and produces less control over the development process. The core services offered by PaaS are as follows:

    • Development tools—that are necessary for software development which includes a code editor, a debugger, a compiler and other supported tools which may offer by other vendors sometimes; but it creates a framework to perform all activities.
    • Middle-layer—as every architecture contains a middle layer that supports all kinds of intermediate activities such that other developers don’t have to create it and can work independently. This lies between user applications and machine’s operating system; like it supports software to take input from keyboard and mouse. It is required to run applications but end user don’t interact it ever.
    • Operating system—All kinds of applications are able to run and develop here through specified vendor.PaaS providers maintain a database with administer in control. They provide a developer with a Database Management System (DBMS).
    • Infrastructure setup—PaaS layer stays above the IaaS layer in cloud service model and contains all the things that is included in IaaS. A PaaS provider either manages storage, servers and data centers.
  3. c) SaaS is a third party hosted software distribution model available to customers over the internet. It is similar to application service provider (ASP) and acts as demand computing software delivery models where the service provider delivers to end user of customer’s software. Since SaaS is a software based on demand model, there is single copy of an application that allows customer to access network based activities. The service provider created the software specifically for SaaS distribution. In the network, the source code is same for all customers and any of features are carried out, it is broadcast to all the customers in the network. Depending upon service level agreement (SLA), the data will be stored either locally or in cloud. Application Programming Interfaces (APIs) help to integrate various softwares with SaaS applications.

    For example, any business organization can write its own software tools that can integrate with SaaS providers for offering or receiving services.

    Basically, the use of SaaS comes when any organization or company wants to launch any new kinds of readymade software quickly of any short term collaboration is required. Sometimes, applications to be used on a short term basis or applications need both web and mobile access.

14.6.2 At a Glance: Cloud Services

Table 14.2 shows comparison of IaaS, PaaS and SaaS where Platform as a Service (PaaS and Infrastructure as a Service (IaaS) give more control to the user segment. SaaS does not require additional efforts to utilize the readymade products.

Table 14.2 Differences between cloud services.

  IaaS PaaS SaaS
Who is the User? System Administrator Software Developers End User
What user get? A logical data store to deploy and create platforms for apps and testing opportunities A logical platform and supported tools to create, deploy and test apps Web software to complete business related task
Controllers Servers, Storage and network virtualization Same as IaaS with OS, Middleware Runtime Same as PaaS with Applications Data
User Controls OS, Middleware Runtime, Applications, Data Applications Data

14.7 Cloud Framework

In the context of exponential growth of data, traditional relational systems fail to meet the present day requirements when dealing with big data. So, a large big data analytics platform like Hadoop, Cassandra, and Voldemort come into the picture.

14.7.1 Hadoop

Hadoop, written in Java and is an open source software that supports all kinds of popular services in cloud environment which consists of two parts or components: i) Map & Reduce (Map Reduce) engine & ii) Hadoop Distributed file systems where Map converts one set of data converted into another set by breaking individual elements into key pairs or tuples. The output moves into Reduce segment to combine tuples to form a smaller set of tuples. It follows a mapping between Map and Reduce segment and scale distributing computing on various nodes. On writing applications with Map Reduce, scaled up even up to thousands of nodes in cluster forms is just a configuration change. It works on top of Hadoop to handle such issues. Different Hadoop segments that work on the commodity level hardware makes it easier to collect, stores, process and analyze large amount of data. Figure 14.10 is the Hadoop ecosystem or platform with various modules [8].

Hive is an Apache open source data warehousing system used for querying and analyzing large volumes of datasets stored in Hadoop files where it uses Hive query language (HQL/HiveQL) that is similar to normal SQL but HQL converts SQL queries into Map Reduce executes on Hadoop. The organization of data is done in tables (consists of rows and columns), partitions and as buckets.

Apache uses a high level language (also known as Pig Latin) that works in between declarative and procedural part of SQL and Java to speed up parallel processing and analysis on large volume of information stored in Hadoop platform. Data sets are compiled into Map Reduce for better performance in Hadoop framework or platform.

HBase is a fault tolerant and works better for structured data and data analyst perform well using HBase. It works on the top of Hadoop distributed system and is highly scalable and provides a distributed data store with consistent read and write options. HBase is a full pledged NoSQL database.

Schematic illustration of Hadoop ecosystem modules.

Figure 14.10 Hadoop ecosystem modules.

The Zookeeper works on to avoid any single point failure through cluster of servers. It runs a special set of rules (protocol) that determines the leader node in a particular time. ZooKeeper distributed servers holds the information that would retrieve the client’s applications.

14.7.2 Cassandra

It is a distributed column-oriented database systems developed by Facebook but built on Google’s Big Table and Amazon’s Dynamo platform which is responsible to handle mission critical data with no compromise in performance and scalability. It works with lower fault handling mechanism at the cloud infrastructure or commodity hardware. It works with various data centers to provide lower latency for users of all kinds of regional outages. It is a right choice for no compromising options when we need scalability and high availability. For all kinds of mission critical data, linear scalability with fault-tolerance mechanism makes it possible for cloud commodity hardware or infrastructure a perfect platform. It supports lower latency by replicating across many data centers and provides a peace of mind to users knowing to exist in regional outages [9]. Cassandra holds the columns (which is not pre-defined as was in a relational table) in a single row. Thus, it is a database with multiple rows in the form of a key-value.

14.7.2.1 Features of Cassandra

Fault Tolerant: Data is copied to various nodes for fault tolerance and replicated to many data centers. A node in failure stage is replaced easily with no time gap.

Performant: Apache Cassandra is a variant of NoSQL that uses partition algorithm to decide the node to organize the data. The performance gets improved through streaming process.

Decentralized: Each node in the cluster acts identically with no network overhead.

Proven: Various companies, organizations like CERN, eBay, GitHub, Go Daddy, Instagram and many others use Cassandra that have large volume of active data sets.

Scalable: With this, big companies like Apple’s; deployed 75,000 nodes storing over 10PB data, Netflix (with 2,500 nodes, 420 TB data and handling 1 trillion requests every day), eBay (over 100 nodes and 250, TB data) are supported easily.

Elastic: Improved throughput (read and write operations) with no time consuming in adding and removal of nodes without any interruption in applications executions.

Durable: Very less chance of data loss even when entire data canter goes down.

Control Mechanism: Option to choose between synchronous and asynchronous copies of data for update and asynchronous operations are can be optimized for cluster consistency with features like Read Repair and Hinted Handoff.

Support: Cassandra supports third party contract and services.

14.7.3 Voldemort

Voldemort is designed in such a way that data are stored across various nodes which works on multiple storage engines in the same framework. This also integrates fast moving, online storage system with large volume of offline data running on Hadoop system. The main features it includes that data is copied on many servers and each server holds only a copy or subset of the portion of total data. Failure of any server is handled transparently with a support of pluggable serialization that allows rich keys and values. There is no single common point of node failure and works independently with each other [10]. To maximize data integrity, data items are versioned without compromising the availability of the system. To measure performance efficiency on a single node, up to 20k operations per second can be expected depending on various factors like machines, network, disk system and replication factor. When using key-value access, read and write access is restricted.

14.7.3.1 A Comparison With Relational Databases and Benefits

Voldemort does not satisfy the general rules of relational database nor does it support object form of database that maps various object based reference graphs nor does it follow any new abstraction like document-orientation. It is a huge, persistent hash table that supports distribution of data across independent nodes. It provides a number of advantages:

  • There is no need of separate caching tier as it combines in memory caching with the storage system and acts just as fast.
  • Do not require any rebalancing of data as portioning is transparent with cluster expansion.
  • Use of simple API can integrate wide range of applications with scope of replicating and placement of data.
  • Support unit testing against a throw-away in memory storage system without need of a real time cluster.

14.8 Conclusions

In this chapter, we have discussed a systematic flow of big data role in achieving the cloud computing environment to manage, storage and process the data. Here, historical view of big data has been shown with its definition and classification and opened a new window for data analytics. Big data is merged with cloud computing and becomes a leading part of the digital world nowadays with its several challenges. The architectural part discussed with cloud services, created a foundation for all kinds of convergence activities. Also, history of cloud computing with its insight in big data showed a general view of service model. Finally, various cloud-based services with some of its features and framework like Hadoop, Cassandra, Voldemort have been discussed.

14.9 Future Perspective

Big data combined with cloud computing platform has created a vast distributed system called the telecom industry that has opened a new window of significant advantage in the transition towards distributed cloud computing. Large companies are in the process to deliver the best in class application performance to their customers with fully leveraging heterogeneous computing and storage capabilities.

Data centers are becoming an emerging part of networks and hardware accelerators and play a vital component of formerly specified as software-only services. As the importance and use of hardware accelerators will only continue to increase, a significant challenge exists for certain exposures. To counter this in the future, companies in turn are suggesting a virtualization technique for any domain specific accelerators (heterogeneous form of chipsets) for multi-tenant use of specific hardware. Also, using zero touch orchestration for any hardware accelerators to design and assign service instances based on the map of resources and accelerators capabilities with the use of artificial and cognitive technologies to optimize business standards and reduce complexities.

Optimizing cloud services to go against homogenization and centralization is not an easy task. Another future trend would be use of quantum level services at the cloud to speed up analytics with maximum optimization of resources available in the network with no failure options and to stabilize the quantum level fluctuation.

References

1. Tsuchiya, S., Sakumoto, Y., Tsuchimoto, Y., Lee, V., Big Data Processing in Cloud Environments. Sci. Tech. J., Fujitsu, 48, 2, 159–168 2012.

2. Hwang, K., Fox, G., Dongarra, J., Distributed and Cloud Computing, University of Southern California, Elsevier Inc., USA, 2011.

3. Laurila, J.K., Gatica-Perez, D., Aad, I., Blom, J., Bornet, O., Dousse, T.-M.-T.O., Eberle, J., Miettinen, M., The mobile data challenge: Big data for mobile computing research, Workshop on the Nokia Mobile Data Challenge, in: Proceedings of the Conjunction with the 10th International Conferenceon Pervasive Computing, pp. 1–8, 2012.

4. O’Leary, D.E., Artificial intelligence and big data. IEEE Intell. Syst., 28, 96–99, 2013.

5. Antonio, R., MIT Technology Review, 31 October, pp. 1, 2011.

6. Hashem, I.A.T. and Yaqoob, I., The rise of big data on cloud computing: Review and open research issue. Inf. Syst., Elsevier Publications, 47, 98–115, 2015.

7. Platform-as-a-service. cloudflare, January 30, 2020.

8. Ashish, S., Technocents, 24 March, pp. 1, 2014

9. Apache Cassandra, Retrieved from https://cassandra.apache.org/, 2016.

10. Project Voldemort, A distributed database, Retrived from http://www.project-voldemort.com/voldemort/, 2017.

Email:: [email protected]

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
34.228.43.90