On proposing and evaluating a NoSQL document database logical approach

• Pages: 398-417
• Published date: 07 November 2016
• Date: 07 November 2016
• DOI: https://doi.org/10.1108/IJWIS-04-2016-0018
• Author: Cláudio Lima,Ronaldo Santos Mello
• Subject Matter: Information & knowledge management,Information & communications technology,Information systems,Library & information science,Information behaviour & retrieval,Metadata,Internet

On proposing and evaluating a

NoSQL document database

logical approach

Cláudio Lima

Federal University of Santa Catarina, Florianopolis, Brazil, and

Ronaldo Santos Mello

Informatics and Statistics Department,

Federal University of Santa Catarina, Florianopolis, Brazil

Abstract

Purpose – NoSQL databases do not require a default schema associated with the data. Even that, they

are categorized by data models. A model associated with the data can promote better strategies for

persistence and manipulation of data in the target database. Based on this motivation, the purpose of

this paper is to present an approach for logical design of NoSQL document databases that consists a

process that converts a conceptual modeling into efcient logical representations for a NoSQL

document database. The authors also evaluate their approach and demonstrate that the generated

NoSQL logical structures reduce the amount of data items accessed by queries.

Design/methodology/approach – This paper presents an approach for logical design of NoSQL

document database schemas based on a conceptual schema. The authors generate compact and

redundancy-free schemas and dene appropriate representations in a NoSQL document logical model.

The estimated volume of data and workload information can be considered to generate optimized

NoSQL document structures.

Findings – This approach was evaluated through a case study with an experimental evaluation in the

e-commerce application domain. The results demonstrate that the authors’ workload-based conversion

process improves query performance on NoSQL documents by reducing the number of database

accesses.

Originality/value – Unlike related work, the reported approach covers all typical conceptual

constructs, details a conversion process between conceptual schemas and logical representations for

NoSQL document database category and, additionally, considers the estimated database workload to

perform optimizations in the logical structure. An experimental evaluation shows that the proposed

approach is promising.

Keywords Web databases, Extended entity-relationship, NoSQL data modeling,

NoSQL document database, NoSQL logical design, Workload-driven approach

Paper type Research paper

1. Introduction

Applications from several domains, such as Web data management, social networks,

sensor networks, e-commerce and educational evaluation, generate a massive amount of

data every day. It brings several challenges for data management in the cloud, including

how to handle and store these data. NoSQL databases (DBs) are designed to manage

large volumes of data, commonly referred to as Big Data, and a large number of

The current issue and full text archive of this journal is available on Emerald Insight at:

www.emeraldinsight.com/1744-0084.htm

IJWIS

12,4

398

Received 13 April 2016

Accepted 5 May 2016

InternationalJournal of Web

InformationSystems

Vol.12 No. 4, 2016

pp.398-417

©Emerald Group Publishing Limited

1744-0084

DOI 10.1108/IJWIS-04-2016-0018

read-and-write operations, a common feature in modern Web applications (Cattell,

2010).

NoSQL DBs support complex data types, semi-structured or unstructured data, and

although they do not require a default schema associated with the data, they are

categorized by data models (key-value, document, columnar and graph-based)

(McMurtry et al., 2013), demonstrating that their data show some degree of structuring.

In fact, data organization in NoSQL DBs requires signicant design decisions because it

affects important quality requirements such as scalability and performance (Bugiotti

et al., 2014). In addition, the importance of a model associated with the data is related to

the denition of better strategies for persistence and manipulation of such data in the

target DB.

In this context of data modeling, conceptual schemas and ontologies are crucial to

dene data semantics, providing access to them with higher accuracy. A traditional DB

design is a process consisting of three data modeling phases (Batini et al., 1992;Elmasri

and Navathe, 2011): conceptual,logical and physical design. At the conceptual modeling

phase, a schema with information about a domain is represented in a high-level

abstraction model. In the sequence, in the logical modeling phase, the conceptual schema

is transformed into a schema with lower abstraction but suitable to the target DB data

model. This logical design phase, specically for NoSQL document DBs, is the scope of

this paper.

In database literature, support methodologies for the logical design of NoSQL DBs is

still a topic little explored, despite its importance (Atzeni et al., 2013). This paper aims to

contribute to this issue by proposing an approach for the logical design of NoSQL

document DBs. This approach consists a process that converts conceptual modeling for

suitable and efcient logical representations for a NoSQL document DB. We chose

document-oriented DBs because they are an appropriate category for Web applications

or applications that deal with Big Data, once they provide semi-structured data storage

and dynamic queries execution, as well as horizontal scalability and high availability

(Kaur and Rani, 2013).

Our conversion approach for generating NoSQL document logical schemas from

conceptual schemas can consider the expected workload of the target application.

Workload information is provided by the designer in terms of the amount of data

instances estimated for the NoSQL DB, as well as the main operations that will be

performed over these data. This information is used to determine an optimized logical

structuring for the NoSQL DB schema, contributing, in general, to a better access

performance for the application. We also evaluate our approach through a case study

with an experimental evaluation in the e-commerce domain. An existing data set was

redesigned by our approach to compare the number of accesses generated by queries

over the redesigned schema as well as over the schemas generated by our approach. We

demonstrate that our method can improve query performance on NoSQL documents by

reducing the number of access to the NoSQL DB. It highlights that our workload-aware

design process is promising.

The remainder of this paper is organized as follows. Section 2 presents an analysis of

related work. Section 3 gives an overview of our approach for converting conceptual

schemas into NoSQL document logical schemas. Section 4 presents a case study to

evaluate our logical design approach, and Section 5 is dedicated to the conclusion.

399

NoSQL

document

database