Big Data in Cloud Computing Review and Opportunities

International Journal of Computer Science & Information Technology (IJCSIT) Vol 11, No 4, August 2019
DOI: 10.5121/ijcsit.2019.11404 43
BIG DATA IN CLOUD COMPUTING REVIEW AND
OPPORTUNITIES
Manoj Muniswamaiah, Tilak Agerwala and Charles Tappert
Seidenberg School of CSIS, Pace University, White Plains, New York
ABSTRACT
Big Data is used in decision making process to gain useful insights hidden in the data for business and
engineering. At the same time it presents challenges in processing, cloud computing has helped in
advancement of big data by providing computational, networking and storage capacity. This paper presents
the review, opportunities and challenges of transforming big data using cloud computing resources.
KEYWORDS
Big data; cloud computing; analytics; database; data warehouse
1. INTRODUCTION
The volume and information captured from various mobile devices and multimedia by
organizations is increasing every moment and has almost doubled every year. This sheer volume
of data generated can be categorized as structured or unstructured data that cannot be easily
loaded into regular relational databases. This big data requires pre-processing to convert the raw
data into clean data set and made feasible for analysis. Healthcare, finance, engineering, e-
commerce and various scientific fields use these data for analysis and decision making. The
advancement in data science, data storage and cloud computing has allowed for storage and
mining of big data [1].
Cloud computing has resulted in increased parallel processing, scalability, accessibility, data
security, virtualization of resources and integration with data storages. Cloud computing has
eliminated the infrastructure cost required to invest in hardware, facilities, utilities or building
large data centres. Cloud infrastructure scales on demand to support fluctuating workloads which
has resulted in the scalability of data produced and consumed by the big data applications. Cloud
virtualization can create virtual platform of server operating system and storage devices to spawn
multiple machines at the same time. This provides a process to share resources and isolation of
hardware to increase the access, management, analysis and computation of the data [2].
The main objective of this paper is to provide review, opportunities and challenges of big data
applications in cloud computing which requires data to be processed efficiently and also provide
some good design principles.
2. BIG DATA
Data which is huge, difficult to store, manage and analyse through traditional databases is termed
as “Big Data”. It requires a scalable architecture for their efficient storage, manipulation, and
analysis. Such massive volume of data comes from myriad sources: smartphones and social

44
media posts; sensors, such as traffic signals and utility meters; point-of-sale terminals; consumer
wearables such as fit meters and electronic health records. Various technologies are integrated to
discover hidden values from these varied, complex data and transform it into actionable insight,
improved decision making, and competitive advantage. The characteristics of big data are.
1.) Volume – Refers to incredible amount of data generated each second from different sources
such as social media, cell phones, cars, credit cards, M2M sensors, photographs and videos
which would allow users to data mine the hidden information and patterns found in them.
2.) Velocity - Refers to the speed at which data is being generated, transferred, collected and
analysed. Data generated at an ever-accelerating pace must be analysed and the speed of
transmission, and access to the data must remain instantaneous to allow for real-time access
to different applications that are dependent on these data.
3.) Variety – Refers to data generated in different formats either in structured or unstructured
format. Structured data such as name, phone number, address, financials, etc can be
organized within the columns of a database. This type of data is relatively easy to enter,
store, query, and analyse. Unstructured data which contributes to 80% of today’s world data
are more difficult to sort and extract value. Unstructured data include text messages, audio,
blogs, photos, video sequences, social media updates, log files, machine and sensor data.
.
4.) Variability – Refers to the high inconsistency in data flow and its variation during peak
period. The variability is due to multitude of data dimensions resulting from multiple
disparate data types and sources. Variability can also refer to the inconsistent speed at which
big data is ingested into the data stores.
5.) Value – Refers to the hidden value discovered from the data for decision making. Substantial
value can be found in big data, including understanding your customers better, targeting
them accordingly, optimizing processes, and improving machine or business performance.
6.) Veracity - Refers to the quality and reliability of the data source. Its importance is in the
context and the meaning it adds to the analysis. Knowledge of the data's veracity in turn
helps in better understanding the risks associated with analysis and business decisions based
on data set.
7.) Validity - Refers to the accuracy of the data been collected for its intended use. Proper data
governance practices need to be adopted to ensure consistent data quality, common
definitions, and metadata.
8.) Vulnerability – Refers to the security aspects of the data been collected and stored.
9.) Volatility – Refers to how long data is valid and the duration for which it needs to be stored
historically before it is considered irrelevant to the current analysis.
10.) Visualization – Refers to data being made understandable to nontechnical stakeholders and
decision makers. Visualization is the creation of complex graphs that transforms the data
into information, information into insight, insight into knowledge, and knowledge into
advantage for decision making [3].

45
Figure 1: V’s of Big Data
3. BIG DATA CLASSIFICATION
1. Analysis Type - Whether the data is analysed in real time or batch process. Banks use real
time analysis for fraud detection whereas business strategic decisions can make use of batch
process.
2. Processing Methodology - Business requirements determine whether predictive, ad-hoc or
reporting methodology needs to be used.
3. Data Frequency - Determines how much of data is ingested and the rate of its arrival. Data
could be continuous as in real-time feeds and also time series based.
4. Data Type - It could be historical, transactional and real-time such as streams.
5. Data Format - Structured data such as transactions can be stored in relational databases.
Unstructured and semi-structured data can be stored in NoSQL data stores. Formats
determine the kind of data stores to be used to store and process them.
6. Data Source - Determines from where the data is generated like social media, machines or
human generated.
7. Data consumers - List of all users and applications which make use of the processed data [4].

46
Figure 2: Big Data Classification
Big data is classified based upon its source, format, data store, frequency, processing
methodology and analysis types as shown in Figure 2.
4. CLOUD COMPUTING
Cloud computing delivers computing services such as servers, storage, databases, networking,
software, analytics and intelligence over the internet for faster innovation, flexible resources,
heavy computation, parallel data processing and economies of scale. It empowers the
organizations to concentrate on core business by completely abstracting computation, storage and
network resources to workloads as needed and tap into an abundance of prebuilt services. Figure
3 shows the differences between on-premise and cloud services. It shows the services offered by
each computing layer and differences between them [5].

47
Figure 3: Summary of Key Differences
The array of available cloud computing services is vast, but most fall into one of the following
categories:
SaaS: Software as a Service
Software as a service represents the largest cloud market and most commonly used business
option in cloud services. SaaS delivers applications to the users over the internat. Applications
that are delivered through SaaS are maintained by third-party vendors and interfaces are accessed
by the client through the browser. Since most of SaaS applications run directly from a browser, it
eliminates the need for the client to download or install any software. In SaaS vendor manages
applications, runtime, data, middleware, OS, virtualization, servers, storage and networking
which makes it easy for enterprises to streamline their maintenance and support.
PaaS: Platform as a Service
Platform as a Service model provides hardware and software tools over the internet which are
used by developers to build customized applications. PaaS makes the development, testing and
deployment of applications quick, simple and cost-effective. This model allows business to
design and create applications that are integrated into PaaS software components while the
enterprise operations or thirty-party providers manage OS, virtualization, servers, storages,
networking and the PaaS software itself. These applications are scalable and highly available
since they have cloud characteristics.
IaaS: Infrastructure as a Service
Infrastructure as a Service cloud computing model provides self-servicing platform for accessing,
monitoring and managing remote data center infrastructures such as compute, storage and
networking services to organizations through virtualization technology. IaaS users are
responsible for managing applications, data, runtime, middleware, and OS while providers still
manage virtualization, servers, hard drives, storage, and networking. IaaS provides same
capabilities as data centers without having to maintain them physically [6]. Figure 4 represents
the different cloud computing services been offered.
Cloud ManagementUser Management

48
Figure 4: Primary Cloud Computing Services
5. RELATIONSHIP BETWEEN THE CLOUD AND BIG DATA
Big data and Cloud computing are closely related to each other. Big data is more about extracting
value while cloud computing focuses on scalable, elastic, on-demand and pay-per-use self-service
models. Big data requires massive on-demand computation power and distributed storage while
cloud computing seamlessly provides elastic on-demand integrated computer resources, required
storage and computing capacity to analyse big data. Cloud computing also offers the distributed
processing for scalability and expansion through virtual machines to meet the requirements of
exponential data growth. It has resulted in the expansion of analytical platforms that are designed
to address the demands of users, especially large data-driven companies in providing contextual
analysed data out of all the stored information. This has resulted in service providers like
Amazon, Microsoft and Google in offering big data systems in cost efficient manner to capture
data and adding analytics to offer proactive and contextual experiences.
Cloud computing environment is a new, great approach to providing IT-related services. Cloud
computing environment has several providers and user terminals. It includes different kinds of
software and hardware, pay-per-use or subscription-based services offered both through the
internet and in real time. Data is collected using big data tools later it is stored and processed in
cloud. Cloud provides on-demand resources and services for uninterrupted data management. The
most common models for big data analytics are software services such as (SaaS), Platform service
like (PaaS) and Infrastructure service like (IaaS). Recently Cloud analytics and Analytics as a
Service (AaaS) are provided to clients on demand. Analytics as a Service (AaaS) provides
subscription-based data analytics software and procedures through the cloud for a fast and
scalable way to integrate data in semi-structured, unstructured and structured format, transform
and analyse them.
Virtualization is the process of creating a software-based, or virtual representation of virtual
desktop, servers, storage, operating system and network resources. It creates a virtual
environment on an existing server to run desired applications, without interfering with any of the
other services provided by the server or host platform to other users. It is a technique, which
allows to share a single physical instance of a resource or an application among multiple
customers and organizations. It does by assigning a logical name to a physical storage and

49
providing a pointer to that physical resource when demanded. Virtualization enables multiple
workloads, running on multiple physical servers, to be consolidated to run on individual virtual
machines that are hosted on a single physical server. Minimizing the number of physical servers
reduces data centre complexity, capital and administrative costs. This reduces and improves
resource utilization and power consumption as compared to the multi-node setup. Virtual
machines are highly flexible and portable which enables the workloads to be transferred to other
physical servers during unplanned downtime, equipment or application maintenance. Virtualized
big data applications like Hadoop provide benefits which cannot be provided using physical
infrastructure in terms of resources utilization, cost and data management. Virtual data includes
wide range of data sources and improves the data access from heterogeneous environments. It
also enables high-speed data flow over the network for faster data processing.
Information privacy and security are one of the important aspects of big data. In cloud as data is
hosted and processed on the third-party services and infrastructure, maintaining the privacy and
security is a big challenge. Bigger the volume, variety, veracity of big data the higher the risk
associated with its privacy and security. Mobile health has revolutionized health care delivery in
many ways. It has enabled people to maintain their lifestyle, health and fitness, drug reference
and diagnosis through various health care mobile applications. Since everything gets transferred
over the mobile internet, it could pose a major threat to data privacy if the network is not secured
properly or is exposed to intruders. Various new laws and privacy policies, regulations,
safeguards, industry standards and Service level agreements must be maintained between
providers and consumers in order to bring confidence in users and made sure that their data is not
exposed. Security of big data in the cloud is important because data needs to be protected from
malicious intruders treats and how the cloud providers securely maintain huge disk space [7].
Figure 5: Big Data and Cloud Computing
The relationship between big data and cloud computing follows input, processing and output
model as shown in Figure 5. The input is the big data obtained from various data sources such as
cellular and other smart devices in either structured, unstructured or semi-structured format. This
voluminous data is cleaned and then stored using Hadoop or other data stores. The stored data is
in turn processed through cloud computing tools and techniques for providing services.
Processing steps includes all the tasks required to transform input data. Output represents the
value obtained after data is being processed for analysis and visualization [8].

50
Internet of Things (IoT) is one of the common factors between Cloud computing and big data.
Data generated from IoT devices is massive and needs to be analysed in real time. Cloud
providers allow data to be transmitted over internet or via lease lines and maintain them in data
stores. Cloud computing techniques and tools are then used on the big data stored in cloud for
filtering and analysis. It provides a pathway for the data to navigate, store and be analyzed. Cloud
computing provides common platform for IoT and big data. Cloud computing, IoT and Big data
have complementary relationship. IoT is the source of the data and big data is an analytical
technology platform of the data as depicted in the Figure [6].
Figure 6: Overview of IoT, Big Data processing and Cloud Computing
6. CASE STUDIES
There are several case studies of big data on cloud computing.
A. Tweet Mining in Cloud
Noordhuis et al. [9] used cloud computing to gather and analyse tweets. Amazon cloud
infrastructure was used to perform all the computations. Tweets were crawled and later page
ranking algorithm was applied. Page Ranking is used by Google to define the importance of a
web page. On an overview, in Page Ranking Algorithm, if an author of a Web page A links to
another Web page B, then author of Page A is likely to find Page B important. This way, the
number of in-edges determine the importance of a certain page. Hence, more the number of in-
links to a page, the more its importance. The same concept is applied to tweets, where instead of a
web page, a tweet user following the other tweet is considered important. We need to crawl the
Twitter social graph to compute PageRank. The data crawled had nearly 50 million nodes and 1.8
billion edges which is about two-thirds of the Twitter’s estimated user base.
B. Redbus
Redbus is an online travel agency for bus ticket booking in India. Redbus decided to use Google
data infrastructure for data processing and analysis in order to improve customer sales and
management of the ticket booking system. Google BigQuery enabled RedBus to process massive
amounts of booking and inventory data within seconds. Applications that reside on multiple
servers continuously streams customer searches, seat inventory and booking information to
centralized data collection system. This massive data collected is then sent to BigQuery, which
runs complex queries and within seconds provide answers to various analytical queries such as

51
how many times a customer searched for a destination and found very few bus services available,
any technical issues that might arise during booking and notify the necessary team. This whole
infrastructure helped RedBus to fix glitches quickly, minimize lost sales and improve customer
service. [9].
C. Nokia mobile company
Nokia mobile phones are been used by many people for telecommunication. In order to better
understand their user interactions and improve the user experience with their phones, Nokia
gathers large amount of data from mobile phones in petabyte scale. For deriving business decision
strategies and get a holistic view of user interaction, the company implemented Hadoop data
warehouse as an infrastructure that could store this daily petabyte of unstructured data collected
from mobile phones in use, services, log files, and other sources. With this setup, Nokia was also
able to perform complex processing and computation on their massive data and gain more
analytical insights on their user interactions such as ‘Which feature did they go to after this one?’
and ‘Where did they seem to get lost? It has also enabled them to create wide variety of
applications such as 3D digital maps that incorporate traffic models that understand speed
categories, recent speeds on roads, historical traffic models, elevation, ongoing events, video
streams of the world, and more. [9].
7. DATA STORES
Modern databases need to handle large volume and different variety of data formats. They are
expected to deliver extreme performance, provide high input/output operations per second
required to deliver data to analytics tools and scale both horizontally and vertically to handle the
dynamic data growth. Database architects have produced NoSQL and NewSQL as alternatives to
relational database. NoSQL supports structured, semi-structured and unstructured data. It scales
horizontally and provides BASE transactions. NewSQL is a new approach to relational databases
that combines ACID transactions of RDBMSs and horizontal scalability of NoSQL. Below are
characteristics of relational database, NoSQL and NewSQL [10, 11].
Characteristics of Databases
Relational
Database
NoSQL
Database
New
SQL Database
ACID property ✔ ✖ ✔
Analytical and OLTP support ✔ ✖ ✔
Data analysis ✔ ✖ ✔
Requires Schema ✔ ✖ ✖
Data format support
✖ ✔ ✖
Distributed parallel processing ✔ ✔ ✔
Scalability ✖ ✔ ✔
8. HADOOP TOOLS AND TECHNIQUES
Big data applications use various tools and techniques for processing and analyses of the data.
Below table represents some of them [12, 13].

52
Tools/Techniques Description Developed by Written in
HDFS Redundant and Reliable massive
data storage
Introduced by
Google
Java
Map Reduce Distributed data processing
framework
Introduced by
Google
Java
YARN Cluster resource management
framework
Apache Java
Storm Stream based task parallelism Twitter Clojure
Spark Stream based data parallelism Berkeley Scala
Map Reduce Java API. Introduced by
Google
Java
Pig Framework to run
script language Pig Latin
Yahoo Java
Hive SQL-like language HiveQL Facebook Java
HCatalog Relational table view of data in
HDFS
Apache Java
HBase NoSQL column oriented Google’s
BigTable
Java
Casandra NoSQL column oriented Facebook Java
Flume Import/Export unstructure or semi-
structure data into HDFS. Data
ingestion into HDFS.
Apache Java
Sqoop Tool designed for efficiently
transferring bulk structured data
(RDBMS) into HDFS and vies
versa.
Apache Java
Kafka Distributed publish-subscribe
messaging system for data
integration
LinkedIn Scala
Ambari Web based cluster management UI Hortonworks Java
Mahout Library of machine learning
algorithms
Apache Java
Oozie Define collection of jobs with their
execution sequence and schedule
time
Apache Java
Sentry Role based authorization of data
stored on an Apache Hadoop
cluster.
Cloudera Java
Zookeeper Coordination service between
hadoop ecosystems.
Yahoo Java
9. RESEARCH CHALLENGES
Big data are huge data sets that are very complex. The data generated is highly dynamic and this
further adds to its complexity. The raw data must be processed in order to extract value from it.
This gives rise to challenges in processing big data and business issues associated with it. Volume

53
of the data generated worldwide is growing exponentially. Almost all the industries such as
healthcare, automobile, financial, transportation etc rely on this data for improving their business
and strategies. For example, Airlines does millions of transactions per day and have established
data warehouses to store data to take advantages of machine learning techniques to get the insight
of data which would help in the business strategies. Public administration sector also uses
information patterns from data generated from different age levels of population to increase the
productivity. Also, many of the scientific fields have become data driven and probe into the
knowledge discovered from these data.
Cloud computing has been used as a standard solution for handling and processing big data.
Despite all the advantages of integration between big data and cloud computing, there are several
challenges in data transmission, data storage, data transformation, data quality, privacy,
governance [14, 15].
Data Transmission
Data sets are growing exponentially. Along with the size, the frequency at which these real-time
data are transmitted over the communication networks has also increased. Healthcare professions
exchange health information such as high-definition medical images that are transmitted
electronically while some of the scientific applications may have to transmit terabytes of data
files that may take longer to traverse the network. In case of streaming applications, the correct
sequence of the actual data packets is as critical as the transmission speed. Cloud data stores are
used for data storage however, network bandwidth, latency, throughput and security poses
challenges.
Data Acquisition and Storage
Data acquisition is the process of collecting data from disparate sources, filtering, and cleansing
data before it can be stored in any data warehouses or storage systems. While acquiring big data,
the main characteristics that pose a challenge are the sheer volume, greater velocity, variety of the
big data. This demands more adaptable gathering, filtering, and cleaning algorithms that ensure
that data are acquired in more time-efficient manner.
Data once acquired, needs to be stored in big capacity data stores which must provide access to
these data in a reliable way. Currently there are Direct Attached Storage (DAS) and Network
Attached Storage (NAS) storage technologies.
Data Curation
It refers to the active and ongoing management of data through its entire lifecycle from creation
or ingestion to when it is archived or becomes obsolete and is deleted. During this process, data
passes through various phases of transformation to ensure that data is securely stored and is
retrievable. Organizations must invest in right people and provide them with right tools to curate
data. Such an investment in the data curation will lead to better quantification of high-quality
data.
Scalability
Scalability refers to the ability to provide resources to meet business needs in an appropriate way.
It is a planned level of capacity that can grow as needed. It is mainly manual and is static. Most of
the big data systems must be elastic to handle data changes. At the platform level there is vertical
and horizontal scalability. As the number of cloud users and data increases rapidly, it becomes a
challenge to exponentially scale the cloud’s ability in order to provide storage and process too
many individuals who are connected to the cloud at the same time.

54
Elasticity
It refers to the cloud’s ability to reduce operational cost while ensuring optimal performance
regardless of computational workloads. Elasticity accommodates to data load variations using
replication, migration and resizing techniques all in a real-time without service disruption. Most
of these are manual instead being automated.
Availability
Availability refers to on demand availability of the systems to authorized users. One of the key
aspects of cloud providers is to allow users to access one or more data services in short time. As
the business models evolve, it would lead to rising demand for more real time system availability.
Data integrity
Data Integrity refers to modification of data only by the authorized user in order to prevent
misuse. Cloud based applications does allow its users to store and manage their data in cloud data
centres, however these applications must maintain data integrity. Since the users may not be able
to physically access the data, the cloud should provide mechanisms to check for the integrity of
data.
Security and Privacy
Maintaining the security of the data stored in the cloud is very important. Sensitive and personal
information that is kept in the cloud should be defined as being for internal use only, not to be
shared with third parties. This would be a major concern when providing personalized and
location-based services as access to personal information are required to produce relevant results.
Each operation such as transmitting data over network, interconnecting the systems over network
or mapping virtual machines to their respective physical machines must be done in a secured way.
Heterogeneity
Big data is vast and diverse. Cloud computing systems need to deal with different formats
structures, semi-structured and unstructured data coming from various sources. Documents,
photos, audio, videos and other unstructured data can be difficult to search and analyse. Having to
combine all the unstructured data and reconcile it so that it can be used to create reports can be
incredibly difficult in real time.
Data Governance and Compliance
Data governance specifies the exercise of control and authority over the way data needs to be
handled and accountabilities of individuals when achieving business objectives. Data policies
must be defined on the data format that needs to be stored, different constraint models that limits
the access to underlying data. Defining the stable data policies in the face of increasing data size
and demand for faster and better data management technology is not an easy task and its policies
could lead to counter productiveness.
Data Uploading
It refers to the ease with which massive data sets can be uploaded to the cloud. Data is usually
been uploaded through internet. The speed at which data is uploaded in turn depends network
bandwidth and security. This calls for improvement and efficient data uploading algorithms to
minimize upload times and provide a secure way to transfer data onto the cloud.
Data Recovery
It refers to the procedures and techniques by which the data can be reverted to its original state in
scenarios such as data loss due to corruption or virus attack. Since periodic backups of petabytes
of data is time consuming and costlier, it is necessary to identify a subset of data valuable to the

55
organization for backup. If this subset of data is lost or corrupted, it take weeks to rebuild the lost
data at these huge scales and result in more downtime for the users.
Data Visualization
Data Visualization is a quick and easy way to represent complex things graphically for better
intuition and understanding. It needs to recognize various patterns and correlations hidden under
massive data. Structured data can be represented in traditional graphical ways, whereas it is
difficult to visualize high diversity, uncertain semi-structured and unstructured big data in real-
time. In order to cope with such large dynamic data, there is a need for immense parallelization
which is a challenge in visualization [16].
10. BIG DATA BUSINESS CHALLENEGES
Utilities: Power consumption prediction
Utility companies use smart meter to measure gas and electricity consumption. These devices
generate huge volumes of data. A big data infrastructure needs to monitor and analyse power
generation and consumption using smart meters.
Social Network: Sentiment analysis
Social networking companies such as Twitter needs to determine what users are saying and topics
which are trending in order to perform sentiment analysis.
Telecommunication: Predictive analytics
Telecommunication provides need to build churn models which depends on the customer profile
data attributes. Predictive analytics can predict churn by analysing the subscribers calling
patterns.
Customer Service: Call monitor
Call center big data solutions use application logs to improve performance. The log files needs to
be consolidated from different formats before they can be used for analysis.
Banking: Fraud Detection
Banking companies should be able to prevent fraud on a transaction or a user account. Big data
solutions should analyse transactions in real time and provide recommendations for immediate
action and stop fraud.
Retailers: Product recommendation
Retailers can monitor user browsing patterns and history of products purchased and provide a
solution to recommend products based on it. Retailers need to make privacy disclosures to the
users before implementing these applications [4].
11. GOOD PRINCIPLES
Below are some of the good design principles for big data applications
Good Architectural Design
Big data architecture should provide distributed and parallel processing through cloud services.
Each node in the cluster must be independent, equal and parallel to other nodes. There must be no
resources shared among these nodes as it would drastically reduce the scalability. NoSQL can be
used for high performance and faster retrieval of data. Lambda and Kappa architectures can be
used for processing in real-time and batch processing mode.

56
Different Analytical Methods
After collecting data from disparate sources, organizations must decide on the analytical method
that best suits their business interest. Some of the commonly used analytical methods are
predictive analytics, stream analytics, data preparation, fraud detection, sentiment analysis. Once,
the analytical methods are decided, big data applications need to take the advantage of data
mining, machine learning, distributed programming, statistical analysis, in-memory analytics and
visualization techniques offered through cloud.
Use appropriate technique
No one technique can be used to analyse data. Since, most data that flows through is unstructured,
we must use appropriate technology stack to analyse this complex data. For instance, if there is a
need for real-time processing, Apache Spark can be used or if we need batch-processing, then
Hadoop can be used. If data must be stored over multiple data centres, then geographic databases
must be considered. Specific analytical tools that work well with established databases must be
used.
Use in-memory analytics
In-memory database analytics can be used to execute analytics where data resides. It performs all
needed analytical functions at runtime. Since the queries and the data reside in the same server’s
memory it reduces response time.
Distributed data storage for in-memory analytics
The data needs to be partitioned and stored in distributed data stores to take the advantage of in-
memory analytics. Cloud computing infrastructure offers this distributed data storage solutions
which must be adopted.
Coordination between tasks and data is required
To achieve scalability and fault-tolerance coordination between data and its processing tasks is
required. Specialized cluster management frameworks such as Zookeeper can be used [16].
12. CONCLUSION
In the big data era of innovation and competition driven by advancements in cloud computing has
resulted in discovering hidden knowledge from the data. In this paper we have given an overview
of big data applications in cloud computing and its challenges in storing, transformation,
processing data and some good design principles which could lead to further research.
REFERENCES
[1] Konstantinou, I., Angelou, E., Boumpouka, C., Tsoumakos, D., & Koziris, N. (2011, October). On
the elasticity of nosql databases over cloud management platforms. In Proceedings of the 20th
ACM international conference on Information and knowledge management (pp. 2385-2388).
ACM.
[2] Labrinidis, Alexandros, and Hosagrahar V. Jagadish. "Challenges and opportunities with big data."
Proceedings of the VLDB Endowment 5.12 (2012): 2032-2033.
[3] Abadi, D. J. (2009). Data management in the cloud: Limitations and opportunities. IEEE Data
Eng. Bull, 32(1), 3-12.
[4] Luhn, H. P. (1958). A business intelligence system. IBM Journal of Research and Development,
2(4), 314-319

57
[5] Sivarajah, Uthayasankar, et al. "Critical analysis of Big Data challenges and analytical methods."
Journal of Business Research 70 (2017): 263-286.
[6] https://www.bmc.com/blogs/saas-vs-paas-vs-iaas-whats-the-difference-and-how-to-choose/
[7] Kavis, Michael J. Architecting the cloud: design decisions for cloud computing service models
(SaaS, PaaS, and IaaS). John Wiley & Sons, 2014.
[8] https://www.ripublication.com/ijaer17/ijaerv12n17_89.pdf
[9] Sakr, S. & Gaber, M.M., 2014. Large Scale and big data: Processing and Management Auerbach,
ed.
[10] Ji, Changqing, et al. "Big data processing in cloud computing environments." 2012 12th
international symposium on pervasive systems, algorithms and networks. IEEE, 2012.
[11] Han, J., Haihong, E., Le, G., & Du, J. (2011, October). Survey on nosql database. In Pervasive
Computing and Applications (ICPCA), 2011 6th International Conference on (pp. 363-366). IEEE.
[12] Zhang, L. et al., 2013. Moving big data to the cloud. INFOCOM, 2013 Proceedings IEEE,
pp.405–409
[13] Fernández, Alberto, et al. "Big Data with Cloud Computing: an insight on the computing
environment, MapReduce, and programming frameworks." Wiley Interdisciplinary Reviews: Data
Mining and Knowledge Discovery 4.5 (2014): 380-409.
[14] http://acme.able.cs.cmu.edu/pubs/uploads/pdf/IoTBD_2016_10.pdf
[15] Xiaofeng, Meng, and Chi Xiang. "Big data management: concepts, techniques and challenges [J]."
Journal of computer research and development 1.98 (2013): 146-169.
[16] Muniswamaiah, Manoj & Agerwala, Tilak & Tappert, Charles. (2019). Challenges of Big Data
Applications in Cloud Computing. 221-232. 10.5121/csit.2019.90918.

Big Data in Cloud Computing Review and Opportunities

More Related Content

What's hot (17)

Similar to Big Data in Cloud Computing Review and Opportunities (20)

More from AIRCC Publishing Corporation (20)

Recently uploaded (20)

Big Data in Cloud Computing Review and Opportunities