Data Management Layer Design

0 Comment

Data management layer design The key definitions are object persistence which is the selection of storage format and to ensure that optimum performance is attain. Object persistence uses four basic formats they include: Files Object-oriented databases Object-relational databases Relational databases 1. Files Below are types of files Sequential and random access files These are the file access modes that are supported by many programming languages such as istream and ostream. Sequential access files These is the type of access mode that if efficient where data is stored according to given attributes which are then used in retrieving the files. The mode is best suited for data which requires reports. Sequential files can be grouped as either ordered or unordered sequential files. Unordered presents data on a disk where the items are attached at the end. While ordered sequential files have items in a proper order. Larger space is used compared to unordered list where the whole list is copied in the disk. Random access files These are files where data is stored without sorting or organization the information in a given manner. These mode is efficient when one is searching for a file because they are retrieved directly without going through the other files like in sequential access file. With random access files one s not able to easily generate reports because information is not orderly stored. Master files These are files that store information permanently where new files are added at the end of the file. These files can be used for data mining and reporting where patterns and trends are discovered. Such information is vital for business decisions. Transaction files There are files that store transaction of daily activities then update the master files. They can be deleted after the master file has been updated. Audit These are files that hold the information before and after a given event or image. These files are used to audit the change after a fact. History or archive file The files which are normally stored offline with information that may not be needed by the organization anymore. Look-up file These are data that are used as references. It is normally unchanged information such as zip codes. Files are very efficient and fast for small amount data. Files are also part of object oriented programming which are suitable for short term storage. A program should be in place to manipulate the files. Redundancy is another shortcoming of files where multiple records are generated which limited security control only through the operating system. 2. Relational databases This is the most common method of data storage which consists of a table that has row and columns. Columns in rational databases are referred to as fields which represents common data of the entities such as names of the students while a row represents a recorded with complete information about an entity. Each table has unique field called the primary key which uniquely identify each of the entities/record in a table. The term relational is for a fact that tables are joined to form a database, the relationship between one table and the other is formed by having a primary key on one table into another table as foreign key. Information is easy searched using structured query language (SQL) which searches the information then presents as one large table form the database. It is also eay to modify or update information by using the SQL commands. This method of data storage is a commercial technology which has been embraced by many players in the business arena. It can handle large and diverse data needs. The limitations are that there are specific formats of data called data types that relational databases store hence limited to those ones only. Object oriented programming is not supported as well. 3. Object Relational databases Relational databases support specific data formats to overcome these challenge ORDBMS was designed it supports user defined data types. Extended SQL is used to manipulate the data. The important fact is that it is still SQL based hence relational database users do not have to learn another language. It can also handle complex types unlike relational databases. Object oriented is the only perspective in this database and it’s also vendor based where it has not been standardized. 4. Object-oriented databases This database is based on two approaches where it can allow adding of extensions to object oriented language that is already in existence or separate the database management systems. Objects, classes and instances are used in this type of databases. This database supports multimedia applications. There is no impedance mismatch in these database approach with direct support of object oriented and can handle complex data types. These technology is still emerging hence can be risky once implemented. To select the object persistence formats several considerations are made which include: a. The type of data type to be supported if it is complex, ORDBMS is appropriate while atomic data types can be implemented using relational databases. b. Type of application systems transaction or decision support where factors such as speed and types queries are important factors. c. What are the existing storage formats these ensure that the transition time is minimized and reduce the effort of learning the new system. d. Future needs of the organization such as the upcoming services to be offered based on changing trends are important factor to the type of object persistence format to use. Other issues such the cost and security are equally important where they should be put into considerations. References Alan Dennis, Barbara Wixom, and David Tegarden, (2005), Systems Analysis and Design With UML 2.0 An Object-Oriented Approach, Second Edition, Chapter 11, John Wiley Sons, Inc. retrieved from, Del Bimbo, A. (1999), Visual information retrieval. Morgan Kaufmann, San Francisco CA. Yannis Tzitzikas, (2007), U. of Crete, Information Systems Analysis and Design, Persistent Data Management Layer Design (III) chapter 10, Fall, retrieved from,