FRAMEWORK FOR DATA MIGRATION BETWEEN NOSQL

• Main Author: Saputra Wijaya - NIM: 23515039 , Yansyah
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/31698
• Institution: Institut Teknologi Bandung

In today's world of information technology, there are two types of databases, namely relational (SQL) and non-relational (NoSQL). The relational database has become a leading database and has been used for a very long time, but due to the rapid development of technology, relational databases are no longer able to meet data storage needs. The three main things that the relational database cannot fulfill are scalability, distributed, and dynamic schemes, so the non-relational database is the answer to the limitations of relational databases. <br /> <br /> <br /> <br /> <br /> NoSQL stands for Not Only SQL. NoSQL illustrates that when designing a software product, there are more than one storage mechanism that can be used based on data requirements. When this book is written based on nosql-database.org there are more than 225 types of NoSQL databases. NoSQL can be categorized based on how the NoSQL stores and manipulates data. Here are the types of NoSQL namely keyvalues, columns, documents and graphics. Each NoSQL has different roles and functions. Although NoSQL goes beyond SQL in some aspects, NoSQL itself still has drawbacks that NoSQL is still relatively new so that support is available only to the community, NoSQL does not have standardized languages so NoSQL can cause problems during migration. <br /> <br /> <br /> <br /> <br /> In information technology (IT), migration is the process of moving from one operating environment to another. There are several types of migration, one of which is data migration. In general, data migration is the process of transferring data between data storage systems, data formats or computer systems. But in reality, data migration in the NoSQL database not only transfers data from one database to another database, but also requires data adaptation, data structures and models in accordance with the target database, without affecting data correctness and integrity. The reason why data migration between NoSQL is that application developers often have difficulty determining data storage needs. This is due to unpredictable data growth in developing an application. <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> Many studies have tried to solve problems regarding the weaknesses that NoSQL has, especially in terms of migration. Started by Srivastava et al by analyzing the NoSQL database so as to produce an overview of NoSQL along with the features, strengths and limitations of each type of NoSQL. Continued with the main problem of NoSQL, namely migration by Scavuzzo et al, but the migration carried out is still limited to NoSQL database type column. Scavuzzo et al found that although in the same database type there were still some obstacles that occurred. Subsequent research was carried out by Bansel et al, research conducted by Bansel et al solved the problem of migration for 3 types of databases at once, namely columns, graphics, and documents using a self-developed framework. <br /> <br /> <br /> <br /> <br /> In general the actual framework is a real structure, or conceptual that is intended as a support or guide for building something that can be expanded by modifying it as needed, either by adding or deleting existing items. In a computer system, the framework can consist of program recommendations that must be built or used, both in the form of techniques / models / algorithms / functions, how the relationships between items, and so forth. So that we can conclude that the existence of a framework can make it easier for developers to migrate data due to the availability of guidance in carrying out the migration process. <br /> <br /> <br /> <br /> <br /> In this study, the author will design a framework that serves as a guide to migrate cloud-based No-SQL data that complements previous research using NoSQL types, namely key-values, columns, documents and graphics. The reason for adding a keyvalue database is that key-values are a major part of the NoSQL database, so that with the presence of key-values increasing the choice of database switching in an existing framework, key-value is the best solution for the needs of IoT in the future. ahead because the key value database has the simplicity needed by IoT, although IoT itself will continue to grow, especially in the direction of integration and infrastructure.