Stages of database application system development

Q: What are the stages of database application system development?
A: The development of database application systems is a software project. It can be divided into the following stages:
1. Planning
2. Demand Analysis
3. Conceptual Model Design
4. Logical Design
5. Physical Design
6. Programming and debugging
7. Operation and Maintenance.
At present, there is no uniform standard for the division of these stages, and each stage is connected to each other, and it is often necessary to perform backtracking and correction.

In the development process of the database application system, the work result of each stage is to write the corresponding documents. Each stage is based on the results of the previous stage. The entire development project is organized, and methodically carried out.
1. Planning

The main task of planning is to analyze the necessity and feasibility.
On the basis of collecting and sorting out relevant information, it is necessary to determine the relationship between the established database application system and the surrounding areas, and to locate the application system, the size, position, and role of the scale must be comprehensively analyzed and demonstrated.

Define the basic functions of the application system and define the scope of database support. Analyze the data source and data collection methods and scope, study the characteristics of the data structure, estimate the size of the data volume, and establish basic requirements for data processing and business standards.

Plan human resource allocation. The technical business level of management personnel and technical personnel involved in the development and future maintenance of system operations should be put forward, and the quality of end users and operators should be evaluated.

Determine the device configuration scheme. To demonstrate the processing capabilities of computers, networks, and other devices in terms of time and space, sufficient internal and external storage capacity is required, the system's response speed, network transmission, and input/output capabilities must meet application requirements and retain margin. Select the appropriate OS, DBMS, and other software. The device configuration scheme should be comprehensively weighed in terms of usage requirements, system performance, purchase costs and maintenance costs.

Estimate System development, operation, and maintenance costs. Predict expected system benefits.

The development progress plan should be drawn up, and specific arrangements should be made for the current work model to transition to the new system.

The results of the planning stage are detailed feasibility analysis reports and database application system plans. The content should include system positioning and functions, data resources and data processing capabilities, human resource allocation, equipment configuration scheme, development cost estimation, and development progress plan.

The Feasibility Analysis report and database application system plan have been approved and approved, and become the general outline of subsequent development work.

2. Demand Analysis

The requirement analysis can be roughly divided into three steps.

(1) collection of demand information. Generally, the collection of demand information is based on the organization settings and business activities.

(2) analyze and sort the collected information. A Data Flow Diagram (DFD) is a description of business processes and data links in the business. Figure 4.1 shows a simple DFD example.

The Data Dictionary (dd) describes all data in the system in detail.

The data dictionary contains the following parts.

· Data item: it is the atomic unit of data.

· Data Group item: it consists of several data items.

· Data Stream: indicates the input/output data of a data processing process.

· Data Storage: The data to be accessed during processing.

· Data processing process description includes: data processing process name, description, input, output, processing work Abstract, processing frequency, processing data volume, response time requirements, etc.

A data flow chart is not only a requirement analysis tool, but also a result of requirement analysis. Data dictionary is the main result of data collection and analysis.

(3) review of requirement information. Each stage of the development process should be reviewed to check whether all tasks are completed, so as to avoid or correct errors and omissions in the work. Employing experts outside of the project to participate in the review can ensure the quality and objectivity of the review.

Review may lead to backtracking of the development process, and may even be repeated multiple times. However, only when all expectations are met can the work of the demand analysis stage come to an end.

The results of the demand analysis stage are to write a realistic and foreseeable demand specification, and a complete set of detailed data flow charts and data dictionaries are attached.
3. Conceptual Model Design
The conceptual model does not depend on a specific computer system. It is a conceptual structure that purely reflects information requirements.

Modeling is based on the requirement analysis results, and data is often abstracted. Common Data abstraction methods are 'clustering 'and 'generalization '.

The ER method is often used to design conceptual models. Use the designed ERTU and then attach the corresponding manual to serve as the stage result.

The conceptual model design can be completed in three steps.

(1) Design a local Conceptual Model

① Determine the scope of the local Conceptual Model

② Define entity

③ Define contact

④ Confirm attributes

⑤ Draw all the partial ergraphs one by one, with corresponding instructions attached

(2) Design a global conceptual model

To create a global ergraph, follow these steps:

① Determine the public entity type

② Merge local ergraphs

③ Eliminate inconsistencies

④ Optimize the global erdiagram

⑤ Draw a global ERTU with a corresponding description file.

(3) conceptual model review

The conceptual model is reviewed in two parts.

The first part is user review.

The second part is the developer review.

4. Logical Design

The main goal of the logic design stage is to convert the conceptual model to the structural data model supported by the DBMS on a specific computer.

The input elements of the logical design include the concept mode, user requirements, constraints, and the features of the selected DBMS.

The output information of logical design includes the following modes and sub-modes that can be processed by DBMS, application design guide, and physical design guide.

(1) design mode and submode

The Relational Database pattern design can be completed in four steps.

① Establish the initial Relationship Model

② Normalization

③ Model evaluation

④ Correction mode

The final mode and sub-mode are determined through multiple mode evaluations and mode revisions.

Write the logical database structure specification.

(2) programming guide

Based on the designed model and application requirements, plan the application architecture, design the application sketch, specify the data access and data processing features of each application, and provide logical interfaces on the program.

Compile the application design guide.

(3) Compile a physical design guide.

Sort out some important data and documents required for the physical design stage according to the designed mode and application requirements. For example, the data capacity of the database, the data capacity of each link (file), the application processing frequency, Operation Sequence, response speed, LRA and TV of each application, and the recommended program access path. These data and requirements will be directly applied to the physical database design.

Compile the physical design guide.

5. Physical Design

Physical Design is to configure a physical structure that is best suited to the application environment for a given Logical Data Model.

The input elements of physical design include: mode and sub-mode, Physical Design Guide, hardware features, OS and DBMS constraints, and operation requirements.

The output information of the physical design is mainly the structure specification of the physical database. The content includes the physical database structure, storage record format, storage record location allocation, and access methods.

The physical design steps are as follows:

(1) Storage record Structure

Design comprehensive analysis of data storage requirements and application requirements, and design the storage record format.

(2) storage space allocation

There are two principles for Bucket allocation:

① Data with high access frequency should be arranged on fast and random devices as much as possible, and data with low access frequency should be arranged on devices with low access frequency.

② Data with strong mutual dependence should be stored on the same device as much as possible, and should be arranged in the adjacent storage space as much as possible.

In terms of improving system performance, we should rationally allocate the designed storage records as a whole to physical storage areas. Make full use of the characteristics of physical order, and assign different types of storage records to different physical groups.

(3) Design of access methods

An access method consists of a storage structure and a retrieval organization. The storage structure limits the access paths that can be used to access the storage records. The Retrieval organization defines the actual access paths used by each application.

(4) Performance Evaluation of Physical Design

① Query response time

The time elapsed between the query start and Result Display is the query response time. The query response time can be further divided into service time, wait time, and delay time.

During the physical design process, the system performance should be evaluated. Performance evaluation includes time, space, efficiency, and overhead.

⊙ The length of CPU service time and I/O service time depends on the application design.

⊙ CPU queue wait time and I/O queue wait time are affected by computer system jobs.

The designer can control the communication delay of the distributed database system to a limited extent.

② Storage space

A bucket stores programs and data. Programs include running applications, DBMS subprograms, and OS subprograms. Data includes user workspace, DBMS workspace, OS workspace, index buffer, and data buffer.

Buckets are divided into primary and secondary storage spaces. The designer can only control the primary storage space to a limited extent, for example, the allocation of a buffer can be specified. However, designers can effectively control the secondary storage space.

③ Overhead and Efficiency

The following overhead should also be considered in the design. As the overhead increases, the system efficiency will decrease.

⊙ Transaction overhead refers to the time consumed from the beginning of the transaction to the end of the transaction. Updating a transaction requires modifying the index, rewriting the physical block, and performing write verification, which adds additional overhead. The update frequency should be listed as a design consideration.

Report Generation overhead refers to the time from data input to result output. The report generation takes a long time to consume CPU and I/O. Filtering is required to remove unnecessary report generation.

The reorganization of the database is also a large overhead. The two factors, data volume and processing frequency, should be taken into account in the design to avoid or minimize the number of restructured databases.

In the physical design stage, the design, evaluation, and modification process may be repeated multiple times, and a complete physical database structure specification will be obtained.

When creating a database, DBA can configure the database using the tools provided by DBMS based on the physical database structure specification.

When the database is running, the DBA monitors the performance of the database and promptly modifies and optimizes the database according to the guidelines in the physical database structure specification to ensure the efficient operation of the database system.

6. Programming and debugging

After the logical database structure is determined, the application design can be developed in parallel with the physical design.

The program module code is usually first debugged in a simulated environment, and then debugged jointly. The main work of joint debugging is as follows:

(1) Create a database structure

Based on the logic design and physical design results, the Data Language (DDL) provided by DBMS is used to compile the source mode of the database, and the target mode is obtained after compilation, the actual database structure can be established by executing the target mode.

(2) debugging and running

After the database structure is created, load the test data so that the database enters the debugging stage. Run the application and Test

(3) load the actual initial data

Before the database is officially put into operation, do the following:

(1) formulate feasible solutions for database reorganization.

(2) develop fault recovery specifications

(3) formulate system security regulations

7. Operation and Maintenance

After the database is officially put into operation, the main tasks in the operation and maintenance phase are:

(1) maintain the security and integrity of the database.

In accordance with the established security specifications and fault recovery specifications, adjust the authorization and change the password in a timely manner in case of system security problems. Promptly detects errors during system operation and promptly modifies them to ensure normal system operation. Take database backup and dumping as daily work. If a fault occurs, use the latest database backup to restore it immediately.

(2) monitor system performance.

The performance monitoring and analysis tools provided by DBMS are used to continuously monitor the running status of the system. When the performance of the database's storage space or response time declines, immediately analyze and study the causes, and take timely measures to improve the performance. For example, you can maintain efficient and normal operation of the database system by modifying some parameters, organizing fragments, adjusting the storage structure, or reorganizing the database.

(3) expand system functions

Based on maintaining the functions and performance of the original system, the system can adapt to changes in the environment and requirements, and expand the original system based on users' reasonable opinions to add new functions.