Understanding the structure of the tables

Each data source (MVS, UNIX, VM, Microsoft Windows) contains unique table features but there are also generalities that apply across all data sources. The same is true of measurement, summary, and prediction data tables. While the summary and prediction tables differ in some respects from measurement data tables, the concept is the same. The following sections use Distributed Systems data as the example but you can apply the general rules across databases.

The tables are joined at several places to construct queries that you run against the tables to generate reports.

Overview

For each set of metrics there are three tables:

• A static table
• A dynamic table
• An interval table

The interval and static tables must have a field called INDEXX. The INDEXX field holds an integral value unique to each record (within a table). The INDEXX field identifies the static object or the interval. The dynamic tables refer to static objects and interval by the INDEXX value.

Static table

The static table contains a list of the objects on which the metrics are measured. An example is:

CAX{MG}S

C = measurement, or collected, data

AX = Distributed systems data

MG = metric group

S = static

For example, the static table CAXCPUS contains a list of CPUs and their associated metrics. A static table might contain nodes, processors, disks, or printers. By convention a static table name ends in S.

Dynamic table

The dynamic table contains the detail ed information about the objects.

CAX{MG}D

C = measurement, or collected, data

AX = Distributed Systems data

MG = metric group

D = dynamic

A dynamic table row associates measurement or summary data with ONE interval and one or more static objects. A dynamic table might contain disk speed, disk utilization, disk size. By convention, dynamic table names end with D.

Interval table

The interval table specifies the details of the time at which the data was collected. It is the key to data selection. To make sure you select the proper interval data, you must understand interval table definitions, including:

• Interval type
• Summary interval selection
• Interval date selection

There is one interval table per data source (Distributed Systems, MVS, VM, and so on) regardless of whether it is collected, MASF, or Exception data.

Interval type

The interval types let you select the type of data to include in the SQL query designed to draw the report. The interval table can contain one of three interval types.

Summary interval selection

Summary records, when requested, are further broken down by sub-type:

Not all subtypes are present in all databases or intervals. If you summarize data, the summarization specification you make in Automator determines the subtypes that
 are present in the database.

Interval date selection

Once the interval type has been defined you can refine your search by date:

Exceptions to the table structure

Exceptions to the UNIX basic table structure include:

• CAXDISK and CAXNODE do not append *S for static tables but do append the *D suffix for the dynamic tables.
• In rare cases, the metric group name in the static table is different from the metric group name in the dynamic name.
• In some cases there is no corresponding static table which usually indicates a link back to the CAXNODE table.

Distributed system measurement tables common to UNIX and Microsoft Windows

The following table lists the root Visualizer table names (minus the S or D designation) that are common to UNIX and Microsoft Windows:

Distributed system measurement tables unique to UNIX

The following table lists the root Visualizer table names (minus the S or D designation) that are unique to UNIX:

Distributed system measurement tables unique to Microsoft Windows

The following table lists the root Visualizer table names (minus the S or D designations) that are unique to Microsoft Windows: