24 - Data Administration

Organizations need to manage data because it can have a direct impact on their performance. It is a critical resource because it helps an organization remember the past, handle the present, and prepare for the future. It is just as important to manage data as it is to manage financial assets. If data are not managed, they may be stored in redundantly across the organization which causes more administrative overhead and use of resources than if the data were managed and held in one consistent manner. Furthermore, the data may not be as accurate when they are not managed than when they are managed. This could result in low quality decisions.

Poor data management can lead to several problems including:

• the same data may be represented through multiple inconsistent definitions
• there may be inconsistencies among different representations
• essential data may be missing from the database
• data may be inaccurate or incomplete
• some data may never be entered into the database, and thus are effectively lost to the organization
• there may be no way of knowing how to locate data when they are needed

The data dictionary contains information about all of the databases, their components, and detailed descriptions of the components, data validation, authorized users, and ownership details. Its purpose is to help database administrators and their staff to manage databases.

The data dictionary should be part of a DBMS because it is an integral component of the DBMS. The data dictionary should automatically reflect the changes that are made in a database. For this to be the case, a closely integrated data dictionary is necessary.

Frequently, external databases are not under the direct control of the organization that is utilizing the data. It is typically a database that is maintained by a separate entity that sells data to interested organizations. The organization that requires access to the data often finds it helpful to extract the data from the external database and to place the data into its internal database(s) to manipulate it.

System level data administration activities deal with management issues applicable to the entire database environment and all databases managed within that environment. It can include planning, development of data standards and policies, establishing integrity procedures, resolving data conflict issues, and managing the DBMS and data dictionary. The system level of data administration is broad in scope and impacts the organizational level. Project level data administration activities support the development and use of a specific database system. Data administration project level activities include defining user requirements, developing data models, training and consulting, establishing and monitoring data integrity, monitoring database usage, and controlling database changes. Project level activities deal with the specifics of database operations.

Data stewards coordinate the management of the data for which they are responsible. For instance, a data steward may be responsible for data definition, quality control and improvement, security, and authorizing access to the database.

Workgroup level databases are shared by people typically in the same work group. They require more planning than the personal level databases but not always as much planning as the organizational level databases. The organizational level database is the most complex form of database in terms of both structure and need for administration. It serves the largest number of people of the three levels of databases (personal, work group, and organizational) and may have a major impact on the organization.

An information architecture is the infrastructure to collect, store, proccess, and disseminate data and information within an organization. It includes details of databases and other organizational memory technologies.

In order to ensure compatibility across the organization, organizational data standards should be devised. These standards define the meaning of the data fields, their size and format. Examples of data items that need to be standardized include sales date, purchase date, customer, and identification number.

To estimate the size of the database, one can measure the current size, discuss the need for expansion, and track the rate of growth over a period of time to benchmark it. Benchmarking should also be used to track the average number of transactions per second for each project or application. This information can be used to determine where the bottlenecks are in the data management system.

An RFP, Request For Proposal, is used to invite a group of vendors to present their products and services that can meet an organization's needs. The RFP outlines the various DBMS requirements the organization is looking for and the desirable characteristics the organization would like its DBMS to have.

The database developer develops the data model and implements the database. This persons tasks include planning, requirements definition, database design, database testing, and database implementation. The data administrator's responsibilities include implementing and controlling the evolution of the database. The data administrator may also perform the consulting tasks required during database development. Data administration is a broader role.

The user plays a critical role in the development of the database. An argument can be made that all of their roles are critical in the development of the database. If only one could be selected, it would perhaps be providing the information and requirements necessary for the design of a high-fidelity data model. Without high quality initial steps in database development, the database may not become a reality.

A medium-sized company that is establishing a data administration group within an MIS Department needs to invest in several software packages. At a minimum, the group should invest in data dictionary software and DBMS software. If the group can purchase additional software, performance monitoring tools, CASE tools, and groupware would be beneficial.

Stakeholders are individuals who stand to benefit from a database project or can influence its success. Their input is valuable in the development of the database. They can lend insights into a variety of issues the database team may face. They may also throw up road blocks for the development team if they do not feel as though their voices are being heard.

CASE tools support project planning, requirements definition, database design, and database testing. They can be used to provide automated assistance for systems development, maintenance, and project management activities.

Groupware is used to enhance the communication between the various groups of people involved in the creation, implementation, and maintenance of a database. These groups include management, users, developers, and computer operations. Its greatest benefits tend to be realized during the requirements definition and database design stages.

This organization could benefit from establishing a decentralized distributed database. The data could be located in various locations, connected via local or wide area networks with several persons being assigned data administration roles and responsibilities. These individuals could be located in the regional offices and be directly responsible for the local data. Data administration also needs to be decentralized so it matches the power structure of the organization.

Both the successful data administration manager and the database technical advisor need to possess interpersonal and communications skills. The data administration manager needs to have keen managerial skills, whereas the database technical advisor needs to have greater command over technical aspects.

NEEDED: DATA ADMINISTRATOR AT A FORWARD THINKING UNIVERSITY

Experience in the design, implementation and maintenance of databases a plus. Sound communication and interpersonal skills a must. Must be a team player. Benefits and salary competitive. Call Erma 919-555-1212

Organizations that find data administration to be highly critical to their mission have the data administration manager report directly to the CIO. The data administration group that reports directly to the CIO tends to focus on the system level data administration issues.

This answer will very from student to student. See Figure 19-2 and Table 19-4 for an outline of the seven database development phases.

It would be most unusual for the development of the database and applications to proceed independently. This could happen only if the database design was extremely well-defined and stable. Under these circumstances, development might then proceed independently if, for example, application programming were outsourced to independent contractors.

Database monitoring is important because it alerts the data administrator to potential problems in the database and is helpful in the projection of future needs for physical storage space. It also allows the data administrator to monitor the database access patterns. The data that may be useful on the database monitoring report would include information about the type of functions requested, by whom and how often, the response time necessary of various functions, the number of disk accesses necessary, the identification of error conditions, and the growth rate of the database.