ADAPTING A NOSQL QUERY LANGUAGE FOR QUERYING SQL AND NOSQL DATABASES

ABSTRACT

Not all kinds of data can find efficient storage and manipulation in Relational (or SQL) database and neither is NoSQL database the best fit for all kinds of data. A hybrid database (a combination of both SQL and NoSQL databases for storage) is a better alternative where structured data are kept in the relational database and the rest in NoSQL database. The hybrid database comes with its challenges; among them is the necessity for the database administrator to learn the query languages of the databases that constitute the hybrid database. This research is focused on using one query language to query the hybrid database via a software layer. NoSQL (MongoDB) query language is adopted as the query language for the hybrid database in this research because it is the fastest growing query language and it is less vulnerable to injections as compared to SQL. The scope of operations supported by the software layer is limited to Create, Read, Update and Delete (CRUD) and MongoDB syntax had to be extended to cater for SQL functionalities necessary to execute CRUD operations in relational databases. The software layer was developed using Java Programming Language; it translates the MongoDB query language syntax to SQL for execution. For evaluation, both databases were fed with the same data via the software layer simultaneously and the CRUD operations were tested simultaneously on the hybrid database and the same result sets were obtained for each database; this asserts that the translation and execution were successful.

CHAPTER ONE

INTRODUCTION

1.1 Background to the Study

Relational database has dominated the database market and had catered for organizational data management needs for decades. It was almost the default choice for any serious data storage in most enterprises. The contemplation was which relational database is to be adopted for a project and not what kind of database will better suit the data in question. The relational database grew and matured over time that it is still considered dependable and reliable (Sadalage et al, 2012). Lately, enterprises have acquired vast volumes of data to store and make available at a very high operation rate to a heavy traffic of users posing a demand that makes the relational database inefficient. This demand can be attributed to the advent of the Web 2.0 that gave rise to very busy websites like social network, e-commerce and the likes. The relational database cannot meet the demands of the new trend. Nance (2013) stated that “relational databases are not well suited for modern web applications that can support millions of concurrent users by spreading the load across a collection of application servers”. This challenge made the category of database called Not Only SQL (NoSQL) to emerge and was easily adopted to cater for the prevalent data storage needs. The NoSQL database found express acceptance but could not become a replacement of the relational database. They have some weaknesses that make the relational database outstanding by the solutions they proffered. Nance et al., (2013) explained that NoSQL databases can be used with applications that have: large transaction volumes, have the need for low-latency access to massive datasets, and have the need for nearly perfect service availability while operating in an 2 unreliable environment. This implies that both types of database (relational and NoSQL) have their strengths and weakness; none is a perfect fit for all kinds of data. There are situations where an enterprise has part of its data suitable for relational database storage and the other part that will fit well into the NoSQL database. Forcing the entire data into a single database will not serve the data management needs efficiently. It can only be better when the data are segregated according to the data model that best suits their usage and are stored in the database system that supports the data models. The application (software) of the enterprise will have to interact with more than a single database which Jafarpour et al., (2015) have termed it as hybrid database.

1.2 Problem Statement

The hybrid database has its demerits. The most crucial problem to this research is that the hybrid database requires the programmer(s) to learn the query languages for each database system that make up the hybrid database; for the programmer to be able to access and manipulate the databases that make up the hybrid database.

Department: Computer Science (M.Sc)

Format: MS Word

Chapters: 1 - 5, Preliminary Pages, Abstract, References

No. of Pages: 67

NB: The Complete Thesis is well written and ready to use.

Price: 20,000 NGN

In Stock

Buy Now

M.SC Thesis in Computer Science