International Journal of Trend in Research and Development, Volume 4(3), ISSN: 2394-9333

www.ijtrd.com

Study of NoSQL Data Types with Big Data

Sarika Rameshwar Rathi,
Computer Engineering Department, MGM‟s Polytechnic, Aurangabad, Maharastra, India
Abstract: The term „BigData‟ explains superior techniques and In such case, new database NoSql plays very significant role in
technologies to capture, store, distribute, manage and analyze big data Analytics. This paper‟s first part explains the concept of
petabyte - or terabytes data with 3 Vs. Data is produced from BigData, second part explains requirement of Nosql concept and
numerous dissimilar resources and can enter in the system at Nosql database last part explains major differences between SQL
various rates. Big data is practical for data sets where their and NoSql.
volume or style is away from the potential of traditional relational
A. Tasks for Traditional RDBMS
databases for capturing, organizing and handing out the data with
low-latency. Relational databases were not intended to handle 1. Relational Database Management System uses Schema, in
with the scale and quickness challenges that face contemporary schema structure of data must be predefined which is
applications, nor were they built to take benefit of the commodity probable to make changes update later, but large changes
storage and handing out power accessible today. NOSQL may be complicated. Now a day‟s those data generated it
provides a mechanism for storage space and recovery of data does not have predefined format, so every time not probable
which is mock-up in means other than the tabular relations used to deal with such complications.
in relational databases. These databases are gradually more used 2. For SQL it is not possible to process random and
in big data and real-time web applications. unstructured data, now a day‟s those data is generated
which is highly flexible ie changing every time.
Keywords: BigData, SQL,Nosql, RDBMS, MongoDB, CouchDB,
3. With Relational Database Management systems, built in
Redis, Apache HBase , Google’s BigTable, Apache Cassandra.
clustering is difficult due to ACID properties.
I. INTRODUCTION 4. With continuously increasing size of database and users
SQL can‟t easily manage those things.
BigData is a gathering of very large amount of datasets. It
5. With development of IT world , new changes are appeared
consists of structured data as relational data, semi structured data
in network database application, such as data model,
as XML data & unstructured data as Word, PDF, Text, Media
distributed architecture and data storage. Also data
logs. Combination of all these contains a huge amount of
extensions, reading and writing speed which is not possible
information. Big data is collection of complex and large data sets,
in SQL.
which include information, may be produced by multiple services
such as Black Box Data, Social media, Stock exchange, Search II. CONCEPT OF NOSQL
engine, sensors used for climate information, digital pictures,
NoSQL Not Only SQL is stands for non relational storage
traffic, software logs etc. Big data at start characterized as
system. Nosql provides a method for storage and retrieval of data
combination of 3V‟s:1st V indicates gigantic Volume of data , 2nd
which is modeled in means other than the tabular relations used in
V represents Velocity of processing huge data and 3 rd V denotes
relational databases. Such databases gradually more used in big
broad variety of different types of data.
data and real-time web applications.
Data is produced from numerous dissimilar resources and can
In Nosql database, data can be added anywhere, at any time.
enter in the system at various rates. The major mission here is to
There‟s no need to specify a document design or schema. Main
merge numerous data from various systems. Here requirement of
features of Nosql databases are scalability, concurrency control,
databases to be skilled to store and process big data efficiently,
consistency, availability, durability etc.
demand for high performance when reading and writing, so the
traditional relational databases is facing many new challenges. A. Architecture of NoSql
Particularly in huge level and high concurrency applications via
the relational database to store and query dynamic user data have
come into view to be inadequate.

Figure 2: Architecture of NoSQL Data

Figure 1: Enlighten concept of 3V‟s

IJTRD | May-Jun 2017

Available [email protected] 137
 International Journal of Trend in Research and Development, Volume 4(3), ISSN: 2394-9333
www.ijtrd.com
These data stores have been intended to offer higher scalability from one model to another using this database.Such information
and availability than conventional relational databases while can be used on social relations or geographic data. Allegro
supporting a simplified transaction and consistency model. Some Graph, Arango DB, Infinite Graph is an example of Graph Based
examples of NoSQL data stores are: MongoDB , Apache HBase , Store
Google‟s BigTable , and Apache Cassandra.
III. COMPARISIONS OF DIFFERENT NOSQL TYPES
WITH ITS PERFORMANCE
Core properties of NoSQL are : (1) they generally use a
distributed and fault-tolerant architecture where additional Table 1: Performance evulation of NoSql Datatypes
computers can be easily added, (2) data is partitioned among
different computers, (3) data is replicated to provide fault Type of Performa Scalabi Flexibil Comple Function
tolerance, and (4) they often support a simplified NoSql nce lity ity xity ality
transaction/consistency model Key High High High Nil Variable
Value (nil)
B. Nosql database types Store
Four different Core NoSql Systems are as : Key value store, Column High High Modera Low Minimu
Document basedstore, Wide Column basedstore, Graph BasedSt te m
Basedstore. ore
Docume High Variabl High Low Variable
Key value store:-In key value store the key may be synthetic or n e (low)
auto-generated while the value can be String. It fundamentally BasedSt
uses a hash table in which there exist a unique key to provide non ore
ambiguous value and a pointer to a particular item of data. A Graph Variable Variabl High High Graph
bucket is a logical group of keys – but they don‟t physically BasedSt e Theory
group the data. There can be identical keys in different buckets. ore
To read a value we require both the key and the bucket because
the real key is a hash (Bucket+ Key). IV. SOME NOSQL DATABASE EXAMPLES
Some of operations performed in this are Insert, Lookup and A. MongoDB
Delete. The simple model provided by key value allows you to
work very fast and efficient represented in Amazon‟s Dynamo. MongoDB is a free and open-source database. It is non-relational
database with dynamic schema, written in C++.Mongodb is one
Document base store: - The data which is collection of key value of the most popular document based NoSQL database as it stores
pair is compressed as document store quite similar to key value data in JSON like documents. It is currently being used by some
store, but difference is that values stored referred as document. big companies like The New York Times, Craigslist, MTV
So here data is stored in documents. Arbitrary data is in Networks.
structured data format. Generally data format is fixed and
structure is flexible.One key difference between key value and As in SQL, RDBMS consist of table,in NoSql ,MongoDB consist
document base is that document base provides attributes metadata of Collection which is a group of MongoDB documents. A
associated with stored content. Apache CouchDB is an example collection exists within a single database. Collections do not
of document store. It uses JSON to store data. enforce a schema. Documents within a collection can have
different fields. Typically, all documents in a collection are of
Wide Column Based store:- are also identified as Extensible similar or related purpose.
Record Store. It can be seen as key value store with two
dimensional key i.e column key and row key with an ability to B. Some of MongoDB Strength
hold very large numbers of dynamic columns. Since the column Speed: it gives good performance, as all the related data are in
names as well as the record keys are not fixed, and since a record single document which eliminates the join operations.
can have billions of columns, wide column stores can be seen as
two-dimensional key-value stores.It can be further extended by Scalability: It is horizontally scalable i.e. you can reduce the
timestamp as in Google‟sBig Table or keyspaces or domains. So workload by increasing the number of servers in your resource
we can say that wide column stores can have many dimensions pool instead of relying on a stand alone resource.
resulting in a structure of multi dimensions.  Manageable: It is easy to use for both developers and
Accumulo, Cassandra, Druid, HBase is an example of Wide administrators. This also gives the ability to shard
Column Basedstore. database
 Dynamic Schema: Its gives you the flexibility to evolve
Graph Based Stores:- In this type data is represented by graph. your data schema without modifying the existingdata
These are usually used for representing data for high information
about flexible number of interconnections. Graph structures are C. CouchDB
used with edges, nodes and properties which provides index free CouchDB is an open source database developed by Apache
adjancy. This kind of database is designed for data whose software foundation, is a multi master application. CouchDB is a
relations are well represented as a graph consisting of elements document based NoSQL database. The focus is on the ease of
interconnected with a finite number of relations between them. use, embracing the web.It uses JSON, to store data (documents),
The type of data could be social relations, public transport links, java script as its query language to transform the documents, http
road maps or network topologies.Data can be easily transferred

IJTRD | May-Jun 2017

Available [email protected] 138
 International Journal of Trend in Research and Development, Volume 4(3), ISSN: 2394-9333
www.ijtrd.com
protocol for API to access the documents. CouchDB became an consists of n number of rows of data whereas NoSQL
Apache project in 2008. Some of CouchDB strengths: databases are the collection of key-value pair,
documents, graph databases or wide-column stores
 Schema Less:- it have a dynamic schema which make it
which do not have standard schema definitions which it
more flexible, having a form of JSON documents.
needs to adhered to.
 HTTP Query:- Easy access databse documents using web
browser.
 Conflict Resolution:- Automatic conflict detection which
is useful in a distributed database.
 Easy replication:- implementing replication is practically
straight forward.
D. Redis
Redis is an open source, advanced key-value store. It is often
referred to as a data structure server, since the keys can contain
strings, hashes, lists, sets and sorted sets. Redis is written in C. It
is generally used for building high performance, scalable web
applications. It is practically used because of its lightening speed.
It can replicate data to any number of slaves. Some of Redis
strength:
 Exceptionally fast − Redis is very fast and can perform
about 110000 SETs per second, about 81000 GETs per  SQL databases are vertically scalable whereas the
second. NoSQL databases are horizontally scalable.
 Supports rich data types − Redis natively supports most  SQL database examples: MySql, Oracle, Sqlite, Postgres
of the datatypes which already known.This makes it and MS-SQL. NoSQL database examples: MongoDB,
easy to solve a variety of problems. BigTable, Redis, RavenDb, Cassandra, Hbase and
 Operations are atomic − All Redis operations are CouchDb
atomic, which ensures that if two clients concurrently
access, Redis server will receive the updated value. CONCLUSION
 Multi-utility tool − Redis is a multi-utility tool and can NoSql database are attractive main element of database scenery
be used in a number of use cases such as caching, today, and with their number of advantages, they can be practical
messaging-queues. in real world applications. Less cost, easier scalability and open
source features construct easier alternative for lots of companies
D. HBase
come across to integrate in BigData. Now a day‟s major business
HBase is a data model that is similar to Google‟s big table data depends on coputers, so IT persons for particular business
designed to provide quick random access to huge amounts of decide which features are most important for their database and
structured data. It is a distributed column-oriented database built based on this select one of the datavase. The choice between
on top of the Hadoop file system. It is an open-source project and NoSql and Sql depends on the complex business needs of an
is horizontally scalable. It is a part of the Hadoop ecosystem that organization with volume, velocity and variety of data what it
provides random real-time read/write access to data in the posses.
Hadoop File System. Some of HBase strength:
References
 Easy Operations:- Easy to perform basicoperations on
HBase using java. [1] Sarika Rathi “A brief Study of Big Data Analytics using
 Fault tolerance:- It leverages the fault tolerance Apache Pig and Hadoop Distributed File Sysetm”
provided by the Hadoop File System (HDFS). [2] NoSQL Databases by Johannes Zollmann
 Easy Storage:-One can store the data in HDFS either [3] BigData NoSql White paper
directly or through HBase. [4] Exploring the Different Types of NoSQL Databases Part ii
 Random Access:-Data consumer reads/accesses the data [5] NoSQL : A New Horizon in Big Data Amarbir Singh
in HDFS randomly using HBase. Published in IJSRSET
[6] SQL: From Traditional Databases to Big Data
V. MAJOR DIFFERENCE BETWEEN SQL AND NOSQL [7] An introduction to Hbase
 SQL databases are primarily known as Relational https://www.tutorialspoint.com/hbase.html
Databases , whereas NoSQL database are primarily [8] An introduction to Redis
known as non-relational or distributed database. https://www.tutorialspoint.com/Redis.html
 SQL databases have predefined schema whereas NoSQL [9] An introduction to MongoDB
databases have dynamic schema for unstructured data. https://www.tutorialspoint.com/MongoDB.html
 SQL databases are table based databases whereas
Sarika R. Rathi, Computer Department , MGM‟s Polytechnic
NoSQL databases are document based, key-value pairs,
College, Aurangabad,India,8087624333/7798615500.
graph databases or wide-column stores. This means that e-mail:- [email protected]
SQL databases represent data in form of tables which

IJTRD | May-Jun 2017

Available [email protected] 139