Comparative Analysis of Legacy Database Systems P a g e | 1 Legacy Database Systems What is a Database? Data is collection of facts and figures which can be processed to produce information whereas a database is the collection of correlated data. For Example, Name of a student, age, class and the subjects can be counted as data for recording purposes. And the names, telephone numbers, and addresses of the people which have been recorded in an indexed address book or stored on a portable hard drive using software such as Microsoft Access or Excel. This collection of related data having an absolute meaning is a database. Database Managеmеnt Systеm: Database Management System or DBMS refers to the mechanism of storing and reclaiming users’ data with uttermost efficiency along with the security attributes. This type of system allows its users to create their own databases which are apropos to the nature of work they desire. Such database systems are highly customizable and proffers bunch of options. The DBMS is therefore such a general purpose system software that facilitates the processes of defining, constructing, manipulating, and sharing of the databases among diverse users and applications. History of Database: The 1960s was the beginning of database management systems (DBMS) when vendors began marketing computerized logistics technologies for manufacturing and wider laboratory use. Intrusion detection system or IDS was one such archaic network databases that was developed by the computer division known as General Electric. Another database was IMS or Information Management System developed by IBM. This was a navigational database. It used the navigational approach in which the navigation method was used as an approach to find information about different various data. Relational Database: The rise of relational databases began in 1970s’ with the proposal of “relational model of data” by E.F. Codd in 1970 which divides the information into various tables with rows and columns each having its own distinctive key. Considering that all the rows have their own unique keys, the information related to any row can be linked to the information of another row simply by the virtuousness of these unique keys. 1980s’ was the forthcoming of desktop computing in which new databases like Dbase and Lotus 1-2-3 were used. Dbase was one of the top selling computer software in the 1990s’. Post-relational database: From 2000’s till today, the next generation of databases used post-relational database model. They are fast and effective and do not require any fixed table schema. Important examples of such databases are NoSQL (Not only SQL) and NewSQL. P a g e | 2 Comparative Analysis of Legacy Systems The main types of early database systems were based on three main archetypes i.e. 1. Hierarchical Model Systems 2. Network Model Systems 3. Inverted File Systems Hierarchical Databases: As its name entails, the Hierarchical Database Model defines hierarchically-arranged data. These databases were IBM's very first database systems, known as “IMS (Information Management System)” that was released in 1960. A hierarchy is just an arrangement of things called “nodes”, and theses nodes are connected by lines or also known as "branches". You can think of these nodes or branches as a connection to the next level of more specific information. The highest node is known as the root node, and queries must pass through this node on their way down the or throughout the hierarchy. Network Model Database: In 1971, the Conference on Data Systems Languages (CODASYL) officially defined the network model which became generally known as the “CODASYL Data Model”. The very first Network Database Model was the invention of Charles Bachman in 1969 to enhance the existing hierarchical database model. This model was created to increase the flexibility and make hierarchical database less difficult to understand. To change this, Charles Bachman developed this model to allow multiple records to be linkеd to the samе ownеr file creating a “many-to- many” relationship rather than a “one-to-many". The indicated “many-to-many” relationships werе formed bеcausе onе ownеr can bе linkеd to many mеmbеr filеs and vicе vеrsa. To makе thе procеss morе faster, Bachman also crеatеd thеsе rеlationships to providе a relation bеtwееn rеcords using “dirеct pointеrs”, which is thе fastеst mеthod of travelling over the whole network model. Inverted Index File Systems: It is an index data structure that stores a mapping from any given content, i.e. words or numbers, to their respective locations in a database file. Its purpose is to allow fast and full text search, at increased processing cost when a document is added to the current database. Several mainframe-based DBMSs have and are currently using inverted index architectures like ADABAS (Adaptable Data Base System), DATACOM/DB, and Model 204. Object-Oriented Databases: The term "object-oriented database" or OODBMs first appeared in 1985 after the notable research projects that included notable names like Encore-Ob, EXODUS, IRIS, ORION and Zeitgeist. These databases store objects rather than data i.e. integers, strings or real numbers. These databases can store deep hierarchies of objects and complex data types by classifying objects for random information. These databases are designed to work efficiently with OOP (Object-Oriented Programming Languages) such as Delphi, Ruby, Python, Perl, Java, C#, Visual Basic .NET, C++, Objective-C and Smalltalk. P a g e | 3 Comparison of OODBMs with RDBMS: As far as data modeling is pertained, OODBMS deals with objects while RDBMS deals with entities. The important features of OODBMs are: 1. Building up of complex objects from simpler ones by application of constructors that represent collections, attributes and relations, and order. 2. The objects in OODBMs have identities independent of their values so that there is a considerable equality and distinction between any two identical objects that point to exactly the same data structure in allocated memory. 3. The objects can be encapsulated i.e. only the operations specified in the interface can be performed on objects by applications and the interface contains hidden implementations which are inaccessible by the applications. 4. Provides database garbage collection i.e. it helps in keeping external programs from having to track the use of object pointers. Next, the main features of RDMS (Relational Database System) are: 1. Efficient secondary data storage management 2. Data provided to be stored is arranged in tables. 3. Data is indexed for quick retrieval of data. 4. A facility primary key is provided to uniquely identify the desired rows. When to Use Object Databases & When Not To? Object databases should only be used when there are complex data relationships. Such relation comprise a many-to- many object relationship. It is better practice not to use the object databases where there would be fewer joint tables or large volumes of a simple transactional data. Comparison of RDBMS with ORDBMS: For RDBMS, SQL2 standard (ANSI X3H2) is used and for ORDBMS, SQL3 standard is applied. RDBMS use b-tree indexes to speed access to any scalar data whereas to define complex data types in an ORDBMS, a specialized index structure is required for efficient management of data. RDBMS is a mature and old software product i.e. have a diverse variety of extensions and features available while ORDBMS is an immature product and still under development. Comparison of OODBMS with ORDBMS: Both DBMSs support any user-defined ADTs, data structures, object identity & reference types, and inheritance. ORDBMSs support extended form of SQL while OODBMSs support ODL/OQL language systems. Both provide DBMS functionality such as recovery and concurrency control. OODBMSs adds DBMS functionality into a programming language, whereas ORDBMSs relates richer data types into a relational DBMS. Conclusion: The markets of object oriented database will continue to develop, but still they will represent only a fraction of traditional databases. Now-a-days, RDBMSs are mostly recommended in the current databases and are likely to coexist with OODBMSs for a very long time. P a g e | 4 Below is the list of currently used modern and legacy Database systems. The systems are listed by type: relational(R), extended-relational(X), object-relational (OR), and object-oriented (OO), network (N) and hierarchical (H): DBMS Vendor Type Primary Market Access (Jet, MSDE) Microsoft R Desktop Adabas D Software AG R Enterprise Adaptive Server Anywhere Sybase R Mobile/Embedded Adaptive Server Enterprise Sybase R Enterprise Advantage Database Server Extended Systems R Mobile/Enterprise Datacom Computer Associates R Enterprise DB2 Everyplace IBM R Mobile Filemaker FileMaker Inc. R Desktop IDMS Computer Associates R Enterprise Ingres ii Computer Associates R Enterprise Interbase Inprise (Borland) R Open Source MySQL Freeware R Open Source NonStop SQL Tandem R Enterprise Pervasive.SQL 2000 (Btrieve) Pervasive Software R Embedded Pervasive.SQL Workgroup Pervasive Software R Enterprise (Windows 32) Progress Progress Software R Mobile/Embedded Quadbase SQL Server Quadbase Systems, Inc. Relationa Enterprise l R:Base R:Base Technologies Relationa Enterprise l Rdb Oracle R Enterprise Red Brick Informix (Red Brick) R Enterprise (Data Warehousing) SQL Server Microsoft R Enterprise SQLBase Centura Software R Mobile/Embedded SUPRA Cincom R Enterprise Teradata NCR R VLDB (Data Warehousing) YARD-SQL YARD Software Ltd. R Enterprise