The security and integrity of information systems is of vast importance to organizations, and significant resources are often deployed in this area. The overall integrity and security of the database is the joint responsibility of the network administrator and the database administrator.
If the database resides on a network, which is typical, network security is the first line of defense to protect the database and the information stored within it. Network administrators can limit access to the directory or files where the actual data and program files are stored. This denies access to those who would delete, modify, or corrupt these files (Andrews, 2003).
Collecting too much or too little data. Collecting too much information slows down the system, clutters screens with unnecessary fields, and inflates the costs of gathering data. Recording too little data can render the database worthless for compiling reports that can help the business grow; instead, money spent on inputting the little data that is there is wasted.
Poorly conceived data fields. The most common mistake is putting too much information in one field—the computer can only sort by field, not by what is in the field. The best rule of thumb is to put each unique record element—ZIP code, phone number, fax number, address—in its own searchable field.
Try to avoid using personal names as the key identifier of a record or as a link between records. Instead, use numbers, assigning a unique number to each record. Personal names cause problems when two or more people have the same name; additionally, if the name is not entered exactly the same way every time it is used as a link—a middle initial is included in one instance and left out in another, for example—the records will not link properly.
Check the database integrity at least once a month. Corrupted links or other problems can creep in over time. Utility programs are available for this function.
Back up information and store it in a separate location, preferably someplace that is fireproof and waterproof.
Set strict standards that must be followed whenever data is input into the system to ensure consistency. This is especially important if multiple departments will be adding data to the system.
Periodically clean out the database, weeding out records that are inactive or no longer relevant. If you do not want to lose those records permanently, create an archive database and move the records into that file.
A. Declarative Constraints
When we design a new DBMS compo- nent such as a new join operator or a new memory manager, we require synthetic database instances with specific characteristics to test correctness and performance of the new component. For example, to test the code module of a hybrid hash join that handles spills to disk, we might need a database instance with a high skew on the outer join attribute. As another example, to study the interaction of the memory manager and multiple hash join operators, we might need a database instance that has particular ...