Before you begin analyzing your site's data access activity, you should review the following sample data access questions, which can be answered using SiteAnalyzer. Following each question are a few of the important decisions you can make based on SiteAnalyzer's answers. You will see that getting answers to even fairly simple questions can help you reduce system usage costs, improve request performance, reduce administrative efforts, increase IS service levels to end users, and improve end users' use of information. |
In this section: Additional Questions You May Want to Investigate With SiteAnalyzer |
Question 1:
Who is running requests?
Importance:
All service organizations can benefit by initially confirming who their users or customers are.
Question 2:
What are the implications of request results being stored? For example, how much report writing is done via screen display or how much output is to files, spreadsheets, or saved tables?
Importance:
Reports that users produce on a regular basis can be scheduled and automatically distributed. Knowing how much output goes to extract or flat files provides estimates on temporary disk space requirements. Data placed in spreadsheets or saved tables may prove useful to other users.
Question 3:
Which requests run the longest?
Importance:
Long-running requests are typically good candidates for performance-tuning improvements. The reorganization of logic or even the simple addition of an index can save hours of processing time for a single request. Long-running requests can be scheduled to run at off-peak hours to reduce the performance impact on the overall system.
Question 4:
How frequently are certain requests run?
Importance:
The same request run multiple times by different users, each producing the same output, could be executed once with a distribution of the report to the necessary users. Requests that are run regularly by certain users may be scheduled.
Question 5:
Which requests use which tables, views, or columns?
Importance:
Data objects often require column revision or content modification. Knowing which requests use these tables and/or columns provides quantitative information for determining the impact of such revisions and allows for a complete implementation of the necessary changes.
Question 6:
Which columns are often used as predicates, and which predicate combinations are frequently included?
Importance:
Columns that are often used with EQ conditions may be good candidates for indexing. Predicate combinations identify a potential new database of information that can be placed in a warehouse environment or otherwise adapted to provide for faster access of the required information.
Question 7:
Which data is not used at all?
Importance:
Unused information can be removed, thus reducing disk space usage and improving request performance. Evidence of unused data may also indicate that users' needs require better analysis.
Question 8:
How detailed is the use of the data? Are most users summarizing or averaging the data, and on which columns?
Importance:
Information about how data is used can identify columns that should be pre-calculated for faster access or reduced storage requirements.
Question 9:
Which time periods are being accessed? Do users examine only the last five of the ten years you have made available? Do users visit the archive?
Importance:
Information about which data is used allows intelligent decisions to be made concerning which data should be archived for efficient use of system storage.
This section has introduced just a few of the issues that can be addressed using SiteAnalyzer. You can begin your analysis of ad hoc data access at your site by seeking answers to these questions. Once you are comfortable with SiteAnalyzer's capabilities, you can formulate further questions specifically designed to meet your organization's needs. The questions you choose to ask will help you decide how to configure SiteAnalyzer's usage monitoring capabilities and, later, which SiteAnalyzer reports to view.
| Information Builders |