Overview
Estimating storage requirements for a Content Manager OnDemand system begins with understanding and documenting end-user requirements for storing and accessing data.
Before you turn requirements into a storage subsystem to support your system, you must also review the various operational and performance issues. For example, Content Manager OnDemand supports up to 128 index fields for each report. However, users should not need a lot of indexes to locate a specific version of a report or a document within a report. The number of index fields that you define has a direct impact on the amount of disk space that you will need for your database. In addition, the more indexes that you define for a report, the longer it will take to load the report into the system. It is important to work with users and understand their data retrieval requirements. Define only the number of index fields that they need. You might have to balance end-user requirements with disk space, the amount of time required to load a report, and other performance issues.
Maintaining a copy of reports on disk can have a significant impact on the amount of disk storage that you need on your system. Most customers store the latest versions or most frequently accessed reports on disk. You should review how users search for and retrieve information from the reports that you plan to store in Content Manager OnDemand. For example, if most retrievals occur in the first 90 days after a report is generated, then you probably want to store the report on disk for at least that length of time. You should choose a time frame for each report which meets the requirements of your users and also makes the best use of available disk space.
- Storage space for application programs and system software, including the base operating system, the Content Manager OnDemand server software, and any other applications on the IBM® i that are required to run your business.
- Storage space for configuration files and control files.
- Storage space for the Content Manager OnDemand system logging facility.
- Temporary storage space for reports received from other systems. In general, you should plan for enough disk space to hold either the largest single report that you will be loading on the system or the total of several reports that might be staged for loading at the same time, which ever requires the most storage space. In many organizations, most versions of a report are similar in size. However, there might be times when a report is much larger than average. For example, a report generated at the end of the month or the end of the quarter might greatly exceed the average report size.
- Temporary storage space for indexing a report on the Content Manager OnDemand server.
- Temporary storage space for loading a report on the Content Manager OnDemand server.
- Storage space for reports stored on disk. This might be zero,
for reports that do not require disk storage. However, a very large
amount of disk space might be required for reports that must remain
on disk for several months or longer.
Content Manager OnDemand compresses report data before storing it. The compression ratio can have a significant impact on the amount of disk space that you need to store a report. Content Manager OnDemand can achieve up to 30:1 compression on line data reports. However, for reports that contain AFP data or image data that is already compressed, the compression achieved will be much lower.
For sorted transaction data, the examples and calculations that follow assume that Content Manager OnDemand creates one indexed item for each group of 100 pages in a report. The number of pages in a group is a parameter that you can configure when you index a report. The IBM Content Manager OnDemand for i: Common Server Indexing Reference provides move information. - Storage space for the database, which includes Content Manager OnDemand system tables (control information
and objects that you define to Content Manager OnDemand)
and application group tables (index data extracted from reports).
The amount of database space that you should plan for a report is
a factor of the number of documents or sections contained in the report,
the number of index fields that you define for the report, the number
of versions of a report (the frequency with which you load a report
on the system), and how long you need to maintain a report on the
system.
For reports that contain sorted transaction data, Content Manager OnDemand can divide the report into groups of a fixed number of pages and create one index row for each group of pages. For reports that contain logical items, such as statements, and policies, Content Manager OnDemand can create one index row for each logical item in the report. Typically the database space required for indexing sorted transaction data is much less than the database space required for indexing reports that contain logical items. Also, index fields provide fast lookup, but require a significant amount of database space.
- Storage space for the Content Manager OnDemand journal and journal receivers. You should plan for disk space for the Content Manager OnDemand journal and receivers used for commitment control.
- Temporary storage space for server print and fax.
- Temporary storage space for importing migrated indexes from archive media to the database.
- Reports that you plan to store on archive media.
- Backup copies of reports stored on archive media. (For critical applications, some customers require that the system maintain two or more copies of a report on archive media.)
You can replace full optical storage volumes with new empty ones in the optical library as needed, if the availability requirements of your system allow you to do so. For example, you might decide to remove full storage volumes from a library one year after the last time that Content Manager OnDemand wrote report data to the storage volume. You could replace the full storage volumes with newly initialized storage volumes to hold the latest reports stored on the system. That way, the latest versions of a report are always available in the library. However, if you need to keep many years of report data online in the library or you store massive amounts of data in your application groups, then you might need to plan on having several optical libraries for your system.