28
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
Big Data: A Brief investigation on NoSQL Databases
Roshni Bajpayee
MATS School of InformationTechnology
Raipur (C.G)
Sonali Priya Sinha
MATS School of InformationTechnology Raipur (C.G)Vinod Kumar
MATS School of InformationTechnology Raipur (C.G)ABSTRACT
As the usage of information technology has increased in the world, the Data generation from various resources has unexpectedly increased. The technology for handling the vast amount of data has not developed as compared to the data generation. Traditional database systems are unable to handle the increased volume of data due to its volume, Variety, Complexity, variability. To deal with this problem, Hadoop Distributed File System (HDFS) like technology is developed. The data to be processed exists in different format that is why the traditional relational database management System is suitable for the big data. To deal with the unstructured data various database tools have been developed. This paper mainly focuses on the various NoSQL Database tools that are available to deal with different types of data. It also includes a brief comparison between (NTFS and HDFS) and (NoSQL and Traditional Relational Database).
Keywords
NOSQL Database, Big data, Big Data Tools, HDFS, NTFS, Hadoop
1. INTRODUCTION
Big Data is a term large amount Data. Which is important new technologies and architecture? Data capture and access third party very easily such as Facebook, other. Big data is a term for large data sets, more different and complex structure with the difficulties of storing, analyze and visualize for further processes or results. The process of research into large amounts of data to reveal hidden patterns and secret correlations named as big data analytics. This useful information for companies or organizations with the help of gaining richer and deeper insights and getting an advantage over the competition. For this reason, big data implementations need to be analyzed and executed as accurately as possible. Big Data applications have high Volume, high Variety and high Velocity.
Some basic properties associated with big data are as follows:
Variety [13][17][20]
Data being generated is not of one category as it not only contains the traditional data but also the semi structured data from different resources like e-mail, Document, Web Pages, Web Log Files, social media sites, etc.
Volume [13][17][20][22]
This characteristic of big data presents the size of the data generated. In the age of information technology, the data is
continuously increasing nearby it will be from petabytes to zettabytes. Networking site stores large amount of data it is very definitely difficult to be handled using traditional system.
Velocity [13][17][20][22]
Velocity deals with the rate at which the data is coming from various sources. This property is not being confined to the rate of incoming data but also rate at which the data flows.
Figure 1: Characteristics of Big Data
Variability [20]
Variability considers the inconsistencies of the data flow data loads become challenging to be maintained. Especially with the increase in usage of the social networking.
Complexity [20]
It is very important aspect of big data because it is quite an undertaking to link, match, cleanse and transform data across system coming from various sources.
Value [20]
User can run certain queries against the data stored and then user got important results from the filtered data obtained and can also rank it according it the dimensions.
2. DAWN OF NOSQL [1][3][4][7][15]
The name “NoSQL” was primarily used in 1998 by Carlo Strozzi [1] for the Relational database management system, Strozzi No SQL. Although, Strozzi coined the term basically to differentiate to his solution from other RDMBS solutions which make use of SQL [15]. He used the term NoSQL just for the reason that his database did not expose a SQL interface. Now, the term NoSQL (Not Only SQL) [1] has come to express a huge set of databases which do not have29
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
characteristics of conventional relational databases and which are usually not queried with SQL. The term re-energized in the recent years with giant enterprises and companies like Google, Amazon, Apache by their own data storage centers to amass and process large amounts of data as they emerge in their applications and stirring up other vendors to take part in it. The main characteristics of NoSQL databases are horizontal scaling, replicating and partitioning data over several servers. In recent years, different kinds of NoSQL databases have been produced mainly by practitioners and web enterprise to fulfill their particular requirements regarding performance, maintenance, scalability and feature-set. In the presentscenario, our need has changed unlike the some years later we were need. Therefore, currently NoSQL has emerged as a solution for today’s data store requirements and has been a subject of talk and research.
3. COMPARATIVE OVERVIEW
Today, People are living in the periphery of big data where each and every moment the data is increasing unexpectedly. It is the massive amount of data (structured, semi-structured and unstructured) being generated with certain velocity from different variety of sources. Traditional File System (TFS) is unable to handle the big data efficiently; therefore Distributed File System (DFS) is taken as a solution over the TFS. Apache Hadoop Distributed File System (HDFS) is playing an important role in the field of big data where commodity hardware is used as data nodes for processing data. Hadoop principally contains of two parts:
File System (Hadoop Distributed File System)
Programming Paradigm (Map Reduce)
3.1 Hadoop Distributed File System
Hadoop Distributed File System is a File System Developed for keeping huge files with streaming data access patterns, running on clusters on commodity hardware. HDFS block size is much larger than that of traditional file system to reduce the number of disk seeks.
3.2 MapReduce
MapReduce is the programming model which runs in the HDFS environment. It consists of mainly two parts- The Mapper and Reducer. Hence, it performs mainly two types of works MAP Task by Mapper and Reduce task by the Reducer. This part is responsible for executing the program in distributed environment and collecting the aggregated result from different distributed nodes i.e. Commodity hardwares.
Table 1: Comparison between NTFS and HDFS [12] NTFS (New Technology
File System)
HDFS (Hadoop Distributed File System)
Files are stored in local file system
Files are distributed across the cluster machines
Block size lies between 512 B and 64 kB
Block size is fixed to 64 MB by default
1 GB file will be split into 16384 blocks
1 GB file will be split into 16 blocks
With seek time of 0.1ms, 1GB will be accessed in 16s
With seek time of 0.1ms, 1GB will be accessed in 16ms
Used for write many read many access
Used for write once read many access
Data will not be replicated Data is always replicated
Scaling up is not possible Scaling wide and scaling deep is possible
Table 2: Comparison between NoSQL Database (HDFS) and Traditional Relational Database [12][15]
4. SUMMARIZED DATA OF VARIOUS
NOSQL DATABASES AVAILABLE
4.1 Key-value stores [3] [4] [7]
SL NoSQL Database (HDFS) Traditional
Relational Database 1. NoSQL is unstructured way of
storing the data.
RDBMS database completely structured way storing of data. 2. The amount of data stored
does not depend on the Physical memory of the system. It can be scaled horizontally as per the requirement.
The amount of data stored mainly depends on the Physical memory of the system.
3. It can effectively handle million and billion of records
It can Effectively handle few thousands of records
4. It is never advised for transaction management
It is best suited for transaction
management 5. Processing time depends upon
number of cluster machines
The processing time depends on the server machine’s
configuration 6. Availability is preferred over
consistency.
Consistency is preferred over availability
7. It follows CAP theorem. It follows ACID property of transaction 8. It scales horizontally as well
as vertically. It scales better vertically 9. There is no need of normalization. Tables must be normalized.
30
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
Key-value store use the associative array as their basic data model. In this model, data is represented as a collection of key-value pair, such that each possible key appears at most once in the collection. The key-value model is one of the easiest non-trivial data models, and richer data models are often implemented on top of it.
4.1.1 Redis
Redis is a data structure server. It is open-source, network, in-memory, and stores keys with optional durability.
4.1.2Riak
Riak is a distributed NoSQL key-value data store that offers extremely high fault tolerance, availability, operational simplicity and scalability. In addition to the open-source versions, it comes in a supported enterprise version and a cloud storage version that is ideal for cloud computing environments.
Table 3: NoSQL- Key Value Store
[3] [4] [7]
SL Na m e Ini tia l/ S ta b le Re lea se Lice n se La n g u a g e De v elo p er Cha ra cte ri stic 1 Re d is 2 0 0 9 /2 0 1 4 BS D Op en S o u rc e C Salv ato re S an fil ip p o Consistency Partition Tolerance Persistence 2 Riak 200 9 /2 0 1 4 Ap ac h e Op en S o u rc e P ro p rieta ry Erl an g Ba sh o Tec h n o lo g ie s High Availability Partition Tolerance Persistence 3 Ae ro sp ik e 2 0 1 2 /2 0 1 4 AG P L P ro p rieta ry C Ae ro sp ik e Persistence 4 Vo ld em o rt 2 0 0 9 /2 0 1 4 Ap ac h e Op en S o u rc e Ja v a Li n k ed In Consistency High Availability Partition Tolerance 5 M em ca ch e DB 20 0 8 /2 0 1 4 BS D C Da n g a In tera cti v e Consistency Partition Tolerance Persistence 6 Hy p ertab le 2 0 1 0 /2 0 1 3 GPL Op en S o u rc e C+ + Hy p ertab le In c Consistency Partition Tolerance Persistence 7 Hib ari 2 0 1 0 /2 0 1 3 Ap ac h e Op en S o u rc e Erl an g Hib ari d ev elo p
ers Strongly Consistent
Highly available 8 S TS d b W4 .0 2 0 1 1 /2 0 1 4 GPL Op en S o u rc e C# STS S o ft SC Consistency High Availability 9 F o u n d ati o n D B 20 1 3 /2 0 1 4 P ro p rieta ry F lo w, C+ + F o u n d ati o n D B Consistency High Availability Partition Tolerance Persistence 10 Dy n am o DB 2 0 1 2 /2 0 1 3 P ro p rieta ry Ja v a, . NET Am az o n .c o m Consistency High Availability 11 To k y o Ca b in et 2 0 0 7 /2 0 0 9 GPL C FAL Lab s Robust High Availability 12 Be rk eley DB 19 9 4 /2 0 1 4 AG P Lv 3 C Slee p y ca t S o ftwa re , late r Ora cle Co rp o ra ti o n Multi Version Concurrency ACID 13 S ca laris /2 0 1 4 Ap ac h e Li ce n se 2 .0 Erl an g Ap ac h e Concurrency Replication ACID 15 h am ste rd b 2 0 0 5 /2 0 1 4 Ap ac h e C/C++ Ro ck S o li d S QL, To n y Ba in ACID
4.2 Document-oriented databases
Document-oriented databases are one of the main categories of NoSQL databases. Document oriented database is developed for storing, managing and retrieving the document-oriented information. The central concepts of a document-oriented database is that Documents. In contrast to relational database
31
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
in which tuple(Row) is the central concept. Document oriented database system is designed around the abstract notion of “Document”.
4.2.1 MongoDB
MongoDB is a document database that provides high availability, easy scalability, and high performance. A MongoDB deployment hosts a number of databases. A manual: data store holds a set of collections. Documents have dynamic schema. Dynamic schema means that document in the same collection do not need to have the same set of fields or structures, and common fields in a collection’s documents may hold different types of data.
Couchbase Server - Couchbase Server, originally known as Membase, is an open source, distributed (shared-nothing architecture) NoSQL document-oriented database that is optimized for interactive applications.
4.2.2 FatDB
FatDB is the next generations NoSQL databases for Windows that extends database functionality by integrating Map Reduce, a work queue, file management system, high-speed cache, and application services.
4.2.3 ArangoDB
ArangoDB is an open source, multi model database that combines a document store with a graph databases. This combination allows you to model your data with a lot of flexibility.I will show you how ArangoDB is difference from other NoSQL database – from its support for transactions to the powerful query languageAQL.
4.2.4 RavenDB
RavenDB is a transactional, open-source Document Database written in .NET, and offering a flexible data model designed to address requirements coming from real-world systems. RavenDB allows you to build high-performances, low-latency applications quickly and efficiently.
4.2.4 OrientDB
OrientDB is database management system written in Java and it is open source NoSQL. It is a document-based database, but the relationship is managed as in graph databases with direct connections between records. It support less, schema-full and schema-mixed modes.
4.2.5BaseX
BaseX is a light-weight and native XML database management system and XQuery processor, designed and developed as a community project on GitHub. It is specialized in querying, storing, and visualizing large XML documents and collections. BaseX is distributed and platform-independent under a permissive free software license.
Table 4: NoSQL – Document oriented
[3] [4] [7]
SL Na m e Ini tia l/ S ta b le Re lea se Lice n se La n g u a g e De v elo p er Cha ra cte ri stic 1 M o n g o DB 2 0 0 9 /2 0 1 4 1 AG P L Op en S o u rc e C+ + , M o n g o DB In c. Consistency Partition Tolerance Persistence 2 Co u ch b ase S erv er 2 0 1 1 /2 0 1 4 Ap ac h e C/C++ Co u ch b ase , In c. Consistency High Availability Persistence 3 Ap ac h e Ja ck ra b b it 2 0 0 4 /2 0 1 4 Ap ac h e Op en S o u rc e Ja v a Ap ac h e S o ftwa re F o u n d ati o n Consistency High Availability Persistence 4 Co u ch DB 2 0 0 5 /2 0 1 4 Ap ac h e Op en S o u rc e Erl an g Ap ac h e S o ftwa re F o u n d ati o n High Availability Partition Tolerance Persistence 5 F atDB 2 0 1 2 /2 0 1 3 P ro p rieta ry C# Fre e Co m m u n it y Consistency High Availability Partition Tolerance Persistence 6 S imp leD B 2 0 0 7 /2 0 1 3 P ro p rieta ry P y th o n , P erl Am az o n High Availability 8 Ara n g o DB 2 0 1 2 /2 0 1 4 Ap ac h e C# , D, ru b y , p y th o n , Ja v a, P y th o n Ap ac h e Consistency Persistence 9 Ra v en DB 2 0 1 0 /2 0 1 4 AG P L Op en S o u rc e P ro p rieta ry .NET Ra v en DB Consistency High Availability Partition Tolerance Persistence
32
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
10 M ark Lo g ic 2 0 0 3 /2 0 1 1 P ro p rieta ry C+ + M ark Lo g ic Co m m u n it y Consistency High Availability Partition Tolerance Persistence 11 Orie n tDB 2 0 0 1 /2 0 1 0 Ap ac h e Ja v a Orie n t Tec h n o lo g ies LTD Consistency High Availability Partition Tolerance Persistence 12 Ba se X 1 9 8 3 /2 0 1 2 BS D Op en S o u rc e Ja v a Ba se X Tea m High Availability 13 S o lr 2 0 0 4 /2 0 1 4 Ap ac h e Ja v a Ap ac h e S o ftwa re F o u n d ati o n High Availability 14 d jo n d b 2 0 1 2 /2 0 1 4 GPL LGP L Op en S o u rc e P ro p rieta ry C+ + Co m m u n it y Consistency 15 iBo x DB 2 0 1 2 /2 0 1 4 P ersiste n ce Ja v a, c# Co m m u n it y Concurrency Transaction support4.3 Column Store
The column of a distributed database is a NoSQL
Object of the lowest rank in a key space. It is a row (a key-value pair) comprising of three parts.
Unique name: column is referenced by it
Value: The substance of the column. It can contain diverse types, like AsciiType, LongType, TimeUUIDType, and UTF8Type among others.
Timestamp: The system timestamp used to resolve the valid content.
4.3.1Cassandra
It is an open source distributed database management system (DDBMS designed to grip huge amounts of data across many commodity servers, offering high availability with no single chance of failure.
4.3.2 Big Table
It is developed keeping in mind the semi-structured data storage. It is a big map that is indexed by a tuple key, column key, and a timestamp. Each value within the map is an array of bytes that is interpreted by the application. Every interaction of
data to a row is atomic, in spite of of how many dissimilar columns are read or written within that row.
4.3.3 HBase
HBase is written in java. It is developed by Apache Foundation. It offers Big Table like capabilities for Hadoop and runs on the Hadoop Distributed File System. It is non-relational, distributed and Open Source and designed after the Google’s’ Big Table.
4.3.4 BangDB
BangDB is developed with the goal to fast, robust, scalable, reliable and very simple to use database for different data management services required by different applications.MongDB comes in the category of multi-flavored distributed key value NoSql database.
4.3.5 Sedna Xml
Sedna Xml is Open Source and it is XML based database management system,
Sedna is an open source database management system that provides native storage for XML data. The distinguishing architecture decisions working in Sedna are (I) for XML data, the schema-based clustering storage strategy is used (ii) use of layered address space for memory management
Table 5: NoSQL - Column Store
SL Na m e Ini tia l/ S ta b le Re lea se Lice n se La n g u a g e De v elo p er Cha ra cte ri stic 1 Ca ss an d ra 2 0 0 8 /2 0 1 4 Ap ac h e Op en S o u rc e Ja v a Ap ac h e S o ftwa re F o u n d ati o n High Availability Partition Tolerance Persistence 2 Big Tab le 2 0 0 5 /2 0 1 0 P ro p rieta ry GPL v 2 C, C+ + , JAVA AG P L Op en S o u rc e P ro p rieta ry Consistency High Availability Partition Tolerance Persistence 3 HBa se 2 0 1 2 /2 0 1 4 Ap ac h e Op en S o u rc e Ja v a Ap ac h e S o ftwa re F o u n d ati o n Consistency Partition Tolerance Persistence
33
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
4 Ba n g DB 2 0 0 8 /2 0 1 4 BS D C+ + IQLECT Consistency High Availability Partition Tolerance Persistence 5 Ha ze lca st 2 0 0 3 /2 0 1 4 Op en S o u rc e Ja v a Ap ac h e 2 Consistency High Availability Partition Tolerance 6 S ed n a Xm l 2 0 0 3 /2 0 1 4 Ap ac h e Op en S o u rc e C, C+ + Ap ac h e Li ce n se 2 .0 Durability Consistency 7 Ac cu m u lo 2 0 1 3 /2 0 1 4 Ap ac h e Li ce n se 2 .0 Ja v a Ap ac h e S o ftwa re F o u n d ati o n Consistency High Availability Partition Tolerance Persistence 8 clo u d era 2 0 0 4 /2 0 1 4 Ap ac h e So ftw are F o u n d ati o n Ja v a, C, C+ + , P y th o n , an d Ru b y Clo u d era , In c. High Availability 10 M o n etDB 20 0 4 /2 0 1 4 M o n etDB Li ce n se (b ase d o n th e M P L 1 .1 ) M o n etDB Li ce n se M o n etDB De v elo p er Tea m Consistency Concurrency 12 Ap ac h e F li n k (in cu b ati n g ) 2 0 1 3 /2 0 1 4 Ap ac h e L ice n se , Ve rsio n 2 .0 Ja v a, sc ale in cu b at o r-fli n k d ev elo p m en t Scalable Reliable Fast Hadoop Compatible 13 Hy p ertab le 2 0 0 9 /2 0 1 3 GNU Ge n era l P u b lic Li ce n se 2 .0 C+ + Hy p ertab le In c. Concurrency Consistency Replication4.4 Graph Database
The graph database is one of the abstract types of data store. It is based on the graph theory and uses the nodes along with edges to represent and store the data. In graph database each and every element contains a direct to its adjacent elements and no index lookups are necessary.
4.4.1Neo4j
Neo4j is most widely used and liked Database in Graph. It is an open-source graph database, implemented in Java. Neo4j is ACID compliant. It’s basic language is java but has interfaces for many other programming languages like Ruby and Python
Table 6: NoSql-Graph Database
SL Na m e Ini tia l/ S ta b le Re lea se Lice n se La n g u a g e De v elo p er Cha ra cte ri stic 1 Ne o 4 j 2 0 0 7 2 0 0 7 /2 0 1 4 s AG P L GPL Op en S o u rc e Ja v a Ne o Tec h n o lo g y High Availability Partition Tolerance Persistence 2 . F lo ck DB 2 0 1 0 /2 0 1 2 Ap ac h e Li ce n se S ca la, Ja v a, Ru b y Twit ter Multi versioning Concurrency Consistency 3 M ero n y m y 1 9 9 6 /2 0 1 2 Op en S o u rc e C+ + Co m m u n it y De v elo p m e nt High Performance ACID Transation 4 . F il am en t 2 0 1 2 /2 0 1 4 BS D Ja v a F il am en t In c. High Availabity Multi version Concurrency 5 Brig h tstarD B 2 0 0 1 /2 0 1 4 M IT Li ce n se C# GitHu b Flexibility Scalability Performance 6 Wh it eDB # 7 3 1 6 4 2 /20 1 3 GPL v 3 a n d a fre e co m m erc ial li ce n ce C Wh it eDB Tea m Protability Prtsistence Concurrency 7 Big d ata 2 0 1 2 /2 0 1 4 GPL v 2 , ev alu ati o n li ce n se , o r co m m erc ia l li ce n se . Ja v a S y sta p S o ftwa re Co m p an y High Performance High availabity 9 Alleg ro Gra p h 20 0 4 /2 0 1 4 P ro p rieta ry co m m erc ial so ftwa re C# , C, Co m m o n Li sp , Ja v a, P y th o n F ra n z, In c. Atomicity Consistency Isolation Durability
34
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
10 Tri n it y 2 0 1 0 /2 0 1 4 P ro p rieta ry C # , C, X 6 4 As se m b ly M icro so ft Highly concurrency Concsistency 11 Hy p erGra p h DB 20 0 1 /2 0 1 0 LGP L Ja v a Ko b ri x In c. High scalability 12 In fo Gri d 2 0 0 8 /2 0 1 1 / AG P Lv 3 , fre e fo r sm all en ti ti es Ja v a Ne tme sh In c. Light Weight 13 TIT AN 2 0 1 2 /2 0 1 4 Op en S o u rc e with li b er al Ap ac h e 2 Ja v a, Blu ep ri n ts, RES T, Co m m u n it y Consistency High availability Fault tolerance 14 DEX 20 0 8 /2 0 1 4 Du al -li ce n se d Ja v a, . NET, C+ + , Blu ep ri n ts In terfa ce S p arsity Tec h n o lo g ie s High Performance 15 In fin ite Gra p h 2 0 1 0 /2 0 1 4 Ev alu ati o n (EUL A), an d co m m erc ial Ja v a Ob jec ti v it y , In c. Highly Scalable5. CONCLUSION
In the age of information technology, data is a very important to extract the useful information. It is obvious that data exists in different format. The processing of big data is still a challenging task. There is no universal tool which can handle enormous and data of various formats. Document oriented, Key-Value pair, Column and graph type of NoSQL databases are developed to handle this variety of data. The summarized discussion about different NoSQL databases is helpful in selection of suitable NoSQL database.
6. REFERENCES
[1] Strozzi, Carlo: NoSQL – A relational database management system. 2007–2010. http://www.strozzi.it/cgibin/CSA/tw7/I/en_US/nosql/Ho me%20Page.
[2] P. Xiang, R. Hou, and Z. Zhou, “Cache and consistency in nosql,” in Computer Science and Information Technology (ICCSIT), 2010 3rd IEEE International Conference on, vol. 6. IEEE, 2010, pp.117–120.
[3] http://nosql.findthebest.com/ [4] http://nosql-database.org/
[5] G. DeCandia et al., “Dynamo: Amazon’s Highly Available Key-Value Store,” Proc. 21st ACM SIGOPS Symp. Operating Systems Principles (SOSP 07), 2007, pp. 205–220; doi: 10.1145/1294261.1294281.
[6] A. Masudianpour, “An Introduction to Redis Server, An Advanced Key Value Database,” SlideShare, 9 Aug. 2013; www.slideshare.net/masudianpour/redis-25088079.
[7] Jing Han; Haihong, E.; Guan Le; Jian Du, "Survey on NoSQL database," Pervasive Computing and Applications (ICPCA), 2011 6th International Conference on, vol., no., pp.363, 366, 26-28 Oct. 2011. [8] http://wiki.apache.org/hadoop/
[9] http://hadoop.apache.org / [10]http://www.cloudera.com/
[11]http://sortbenchmark.org/YahooHadoop.pdf
[12]Tom White, Hadoop: The Definitive Guide, 3rd Edition, O'Reilly Media, 2010
[13]Stephen Kaisler, Frank Armour, J. Alberto Espinosa, William Money, “Big data: issues and challenges moving forward”, IEEE, 46th Hawaii International Conference on System Sciences, 2013.
[14]Frank, L., "Countermeasures against consistency anomalies in distributed integrated databases with relaxed ACID properties," Innovations in Information Technology (IIT), 2011 International Conference on , vol., no., pp.266,270, 25-27 April 2011.
[15]M. Stonebraker, “Sql databases v. nosql databases,” Communications of the ACM, vol. 53, no. 4, pp. 10–11, 2010.
[16]Sachchidanand Singh, Nirmala Singh, “Big data analytics”, IEEE, International Conference on Communication, Information & Computing Technology (ICCICT),Oct. 19-20, 2012.
[17]Sagiroglu, S.; Sinanc, D., "Big Data: A review," Collaboration Technologies and Systems (CTS), 2013 International Conference on , vol., no., pp.42,47, 20-24 May 2013.
[18]Wielki, J., "Implementation of the big data concept in organizations - possibilities, impediments and challenges," Computer Science and Information Systems (FedCSIS), 2013 Federated Conference on, vol., no., pp.985, 989, 8-11 Sept. 2013.
[19]Segev, A; Chihoon Jung; Sukhwan Jung, "Analysis of technology trends based on big data," Big Data (BigData Congress), 2013 IEEE International Congress on, vol., no., pp.419, 420, June 27 2013-July 2 2013
[20]Katal, A; Wazid, M.; Goudar, R.H., "Big data: issues, challenges, tools and good practices," Contemporary Computing (IC3), 2013 Sixth International Conference on , vol., no., pp.404,409, 8-10 Aug. 2013.
35
Roshni Bajpayee, Sonali Priya Sinha, Vinod Kumar
[21]K. Kambatla, G. Kollias, V. Kumar, A. Grama, “Trends in big data analytics”, J. Parallel Distrib. Comput. (2014), http://dx.doi.org/10.1016/j.jpdc.2014.01.003 [22]Sheth, Amit, "Transforming big data into smart data:
deriving value via harnessing volume, variety, and velocity using semantic techniques and technologies," Data Engineering (ICDE), 2014 IEEE 30th International Conference on , vol., no., pp.2,2, March 31 2014-April 4, 2014.
[23]Saha, Barna; Srivastava, Divesh, "Data quality: the other face of big data," Data Engineering (ICDE), 2014 IEEE 30th International Conference on, vol., no., pp.1294, 1297, March 31 2014-April 4 2014.