You must budget and plan for such benchmarks.Īfter the application is in production, run benchmarks regularly in the production environment and store their results in a database table. Such benchmarks are especially important if you expect the user or transaction load to increase over time. You must develop much larger, more complex benchmarks to measure application performance during peak user loads, peak transaction loads, or both. You can create small benchmarks that measure performance of the most important transactions, compare different solutions to performance problems, and help resolve design issues that could affect performance. For meaningful benchmark results, you must test the application in the environment where you expect it to run. Results from such benchmarks provide a performance baseline for the application. Usually, you first run benchmarks on an isolated single-user system to minimize interference from other factors. Benchmark results either validate application design or raise issues that you can resolve before putting the application into production. If the application does not meet the benchmarks, tune your SQL statements to perform optimally, first with no workload and then with increasing workloads.īenchmarks are tests that measure aspects of application performance. For information about benchmarking your application, see Section 2.1.3. For information about tools for testing performance, see Section 2.2.īenchmark tests determine whether the application performs as expected under various workloads (including peak activity) with simulated real-time operations, such as adding data and users. Select the programming language in which to develop the application, develop the user interface, create and test the transactions, and so on.Įnsure that the application runs to specification.Įnsure that all components are exercised, the application is fully operational, and the database features that the application uses are optimally configured. In a test environment that is as similar as possible to the production environment, run the scripts that implement the physical database design. Implement the application in a test environment. Implement the database application by following this basic procedure: Prototype transactions in SQL and develop a volume table that indicates the size of your database.ĭetermine which tables are accessed by which users in which sequences, which transactions read data, and which transactions write data.ĭetermine whether the application mostly reads data or mostly writes data.Ģ.1.1.3 Implement the Database Application Trace transaction paths through the logical model. Know the most common transactions and those that users consider most important. Normalize the tables to minimize redundancy and dependency. Map relationships among columns and tables, determining primary and foreign key attributes for each table. Translate the data requirements into data items (columns). To model relationships between database objects:
XSECTION 7.7 TRACES OF THE SECTION PLANE UPDATE
Effective logical design considers the requirements of different users who must own, access, and update data. The logical design models both relationships between database objects and transaction activity of the application. The logical design is a graphical representation of the database.
0 kommentar(er)
