CS 523 DATABASE DESIGN CS 523 Database Design

CS. 523 DATABASE DESIGN

CS 523 Database Design Instructor : Somchai Thangsathityangkul You can download lecture note at http: //kbucomsci. weebly. com/ Class Presence 10% 2 Quiz 10% Project 20% Midterm 20% Final 40% Total 100%

Database Design Life Cycle 3

Requirements Analysis �The most important step n implementing 4 a database system is to find out what is needed — What type of a database is required for the business organization, daily volume of the data, how much data needs to be stored in the master files etc. In order to collect all this required information, a database analyst need to spend a lot of time within the business organization talking to people, end users and get acquainted with day-to-day

Database Design �In this stage the database designers 5 will make a decision on the database model that is perfectly suited for the organization’s requirements. The database designers will study the documents prepared by the analysts in the requirements analysis stage and then start developing a system that

Evaluation and Selection �Once the data model is designed, tested and demonstrated, the next phase is to evaluate the diverse database management systems and choose the one that is perfectly suited for the requirements of the organization. In order to identify best performing database for the organization, end user should be involved in this phase. 6

Logical Database Design �Once the evaluation and selection phase is completed successfully, the next step n the database development life cycle is logical database design. The conceptual design is translated into internal model in the logical design phase. This includes the mapping of all objects i. e. tables design, indexes, views, transactions, access privileges etc. 7

Physical Database Design �Physical database design is the procedure of selecting and characterizing the data storage and data access of the database. The data storage depends on the type of devices supported by the hardware, the data access methods and the DBMS. 8

Implementation �In most databases a new database implementation needs the formation of special storage related constructs to house the end user tables. These constructs typically comprise storage group, table spaces, data files, tables etc. 9

Data Loading �Once the database has been created, the data must be loaded into the database. The data required to be converting and migrating to the new database, if the loaded data is currently stored n a different system or in a different format. 10

Testing and Performance Tuning �The next phase is testing and 11 performance tuning, this phase starts soon the data is loaded into the database. In this phase, database is tested and fine -tuned for performance, integrity, access and security constraints. It is very important that the database administrators and application programmers work together during this phase, because testing and performance tuning happens in parallel.

Operation �Once the data is loaded into the 12 database and it s fully tested, the database is than released into production. �In operation phase, the database is accessed by the end users and application programs. This stage includes adding of new data, modifying existing data and deletion of obsolete data. The database administrators perform the administrative tasks periodically such as performance tuning,

Maintenance �Database maintenance phase is very 13 important and it is one of the ongoing phases in DDLC. Factors such as new business needs, new information requirements, acquisition of new data etc will make it essential to formulate ongoing changes and improvements to the existing design. The major tasks in this phase include: database backup and recovery, performance tuning, design modifications, access management and audits, usage

Growth and change �Modify the database to reflect changes in the organization. Monitor database performance for efficiency and customer satisfaction. 14

Database and Application Development Life Cycle

Process Modeling

Logical Data Modeling

Database Design

Database Generation

Approaches to Modeling �Top-down modeling �Bottom-up modeling �Targeted modeling Database Synchronized Models Database

Top-Down Modeling: Designing a New Database 1. Business information 2. One process model (DFD) 3. One logical data model (ERD) 4. One multidimensional model 5. One or more relational models 6. One or more physical models for each relational model Database

Bottom-Up Modeling: Modifying an Existing Database 1. Produce a relational model. 2. Modify the relational model and create additional relational models. 3. Reverse engineer the logical model from the relational model. 4. Modify and check the design rules for the logical model. 5. Generate the modified DDL code. Database

Targeted Modeling � Targeted Modeling: Maintaining Existing Models by Adapting to New Requirements 1. Change the logical data model. 2. Engineer to modify the relational model. Synchronized Models 2. Engineer to modify the logical data model. 1. Change the relational data model. Database

What Is a Process Model? � A process model documents the processes that the business performs and shows how the data flows through processes between external sources and information stores. � The diagram for a process model is called a data flow diagram (DFD).

Why Create a DFD? � Data flow diagrams help to: �Document project boundaries �Understand current area of interest �Find inefficiencies in the current process �Communicate complex ideas and processes with users �Determine data flow identification for data modeling

Validating Your Data Flow Diagram DFD Rules: Process �Every process has to have at least one input and one output. �Use verb phrase labels. �Each process has a unique name.

DFD Rules: External Agents �External agents move directly to and from a process. �Use a noun phrase label.

DFD Rules: Information Store �Data cannot move directly from one information store to another information store. �An information store cannot move directly from or to an external agent. �One end of an information flow must be a process. �Use a noun phrase label.

DFD Rules: Information Flow �An information flow has only one direction of flow between objects (not a double arrow line). �An information flow to an information store creates, updates, or deletes data. �An information flow from an information store retrieves data. �Use a noun phrase label.