There are a number of areas of application development and site maintenance for which SiteAnalyzer can be useful. This section describes some of these areas, which all fall under the category of performance analysis. SiteAnalyzer has several reports that are specifically geared towards these performance analysis issues. |
In this section: Benchmarking and Service Management Summarization and Pre-Aggregation Change Impact: Effects of Application Revision Quantifying Return on Investments |
Most organizations do not consider the performance of the request environment until it has already begun to deteriorate. By this time, the data source design is complete and the programs are in production. Companies must then implement service management, or Service Level Agreements (SLAs), to manage the high-performance production environment at an appropriate, justifiable cost. The application developer can use data in SiteAnalyzer reports and graphs to evaluate conditions in test environments before the investment is made into a viable production environment, and also later to determine if SLAs are being implemented effectively.
As IT, OLTP, and data warehouse systems grow, large volumes of unused (or dormant) data accumulate, which make searches for more critical data inefficient.
There are two categories of dormant data that SiteAnalyzer identifies:
SiteAnalyzer lets you examine the use of date selections and relations in requests at your site to determine what data is largely bypassed by your site's request activity. This data can then be deleted or archived to allow for better performance.
When reports are run multiple times, it is costly and inefficient to repeatedly calculate column totals and averages. A more efficient way would be to calculate column values once and store the results in summary tables for repeated use. This technique is known as pre-aggregation or summarization. SiteAnalyzer's reports can show you which data is most frequently summarized, providing valuable information for building summarization tables.
Although traditional theories about data source design encourage normalization to maintain referential integrity, denormalization can be efficient and resource-cutting when planned and implemented judiciously. Denormalization can, however, increase expenses for updating data as well as increase the amount of redundant data, thus leading to increased disk usage. Nevertheless, with denormalized data sources, you may cut CPU time by 40-50% and see 25-30% faster response times.
Individual departments often need to analyze only a subset of the data found in OLTP application systems and data warehouses. For these purposes, data marts offer better customization and control than dealing with the entire system. Moving data into a data mart on a departmental machine can reduce processing and storage costs as well as allow you greater control over several areas of data source design and maintenance:
SiteAnalyzer lets you focus in on which data, data sources, and columns are most used by particular departments to determine the most effective strategies for designing your data marts.
Think how useful it would be to be able to project how proposed changes will affect your site's activity. SiteAnalyzer helps you assess what end users and applications will be affected by data source schema change. SiteAnalyzer's reports help you identify every user of a data source and column, and pinpoint every field accessed by requests, to judge the overall impact of impending changes.
SiteAnalyzer contributes to capacity planning by helping you determine who "visits" data sources and which applications are most mission-critical to your user community. The quantifiable data supplied by SiteAnalyzer justifies the expenses of data warehouses, new release software, and hardware upgrades.
Information analysis has become a commodity that can be sold to business partners, vendors, customers, and external resellers. SiteAnalyzer's usage monitoring functionality lets you track who is accessing your data sources in order to implement usage-based billing and user chargeback systems.
Educating the end-user community about request preparation is crucial to improving system performance. SiteAnalyzer's usage monitoring capabilities let you track what users are doing wrong so you can best determine what areas they need to be informed about and trained in. SiteAnalyzer can also help you implement a chargeback system at your site to inform individual users about their resource consumption and encourage them to improve their techniques.
When you are considering downsizing and right-sizing, it is crucial that you understand what data is used, how it is used, and who uses it at what times. The usage details supplied by SiteAnalyzer identify the least active data sources, those with the smallest audiences, the least active and seldom-used fields, along with a variety of other information to help you develop a clear road map for data migration.
| Information Builders |