research paper1

Data, we are in a world where everything runs with data. Every company has their own challenges with data irrespective of whether it is a small company or big. The past 25 years, Data has raised in a massive scale in diverse fields including software based medical rehabilitation system [1] and sports coaching [2]. According to the report of International Data Corporation (IDC), the over-all created data in the world will reach 44 ZB or trillion gigabytes during the time of 2013 to 2020 [3]. This data refers to Big-Data. Databases are the one which hold this information and it’s a key component of information systems. Proper use of database can be used for storing and retrieving the data in meaningful manner.

World was running completely fine with Relational databases until Big-data arrived and it’s challenging to analyze this data using traditional database management tools and companies are looking for alternative solutions that includes NOSQL databases. NOSQL database store/process data which does not follow traditional model of relational DBMS.

There are mainly four types of NOSQL database Column Oriented, Key-Value type, Document based or Graph based [4]. The current popular NOSQL databases include Redis, MongoDB, Couch-base, Cassandra, HBase. Based on the company’s requirements there is need to investigate which NOSQL is best when it comes to using NOSQL in the company comparing various NOSQL databases which are available.

NOSQL databases are known to provide easier scalability, storage flexibility, and greater data manipulation and performance improvement. So, any company will try to improve where and when there is room for it, NOSQL provides such environment to dig more on this to figure out which NOSQL approach could be best.

Research questions include 1) Comparing Various NOSQL databases. 2) Which NOSQL is best for big-data processing. 3) Are there any transactions possible in NOSQL databases?

If the best NOSQL database is figured out among the all based on the company requirements, it would eventually lead to better performance in all aspects of the firm.

With the raise in volume of data in every corner, there is a need to store and retrieve the data in more appropriate and efficient way. With the rise of Big-data and web applications showing more complex data is the birth place to NOSQL. As this is fairly new when compared to existing RDBMS which are decades old, there is a need to take a deep dive to look at the various NOSQL databases which are available to us , which is best among them and what they offer.

RDBMS: RDBMS stands for Relational Database Management Systems. RDBMS in this research stands for a model database management system. Relational databases follow relational model invented by Edgar F. Codd that identify three components in a data model that are data structure, operators and integrity rules [5]. Relations, attributes, tuples and keys give structure in RDBMS. Operations such as select, insert, update and delete are common operators in relational databases. Keys (primary) among other different referential integrity rules primarily define integrity rules.

NOSQL: NOSQL systems are mostly non-relational database systems that are distributed and are understood as Not Only SQL. NOSQL databases are known to provide easier scalability, storage flexibility, and greater data manipulation and performance improvement. There are various types of NOSQL database systems among which Key-value stores; Wide-column stores, Graph databases and Document stores are identified most commonly [5]. MongoDB, Cassandra, DynamoDB and CouchDB, Neo4j, Riak are the more popular NOSQL databases used commonly in today’s environment.

The focus of this research solely is to compare various NOSQL databases and to decide which is best among them. However, there is a need to take a deep dive considering on-premises vs cloud once the decision of NOSQL is made.

This Research doesn’t include any tests or experiments and is limited to the existing studies out there in the world that includes small articles to research papers.