ACT Permanent DASD Space Accounting
Accounting for permanent DASD data sets requires a system to identify the pertinent data set names and the space allocated to them. Then data set ownership is ascertained to identify the cost center to be charged. Data set names, standardized by a data set naming convention, are used to determine ownership. VTOC scans usually produce the necessary information.
rm
Accounting for permanent DASD data sets requires a system to identify the pertinent data set names and the space allocated to them. Then data set ownership is ascertained to identify the cost center to be charged. Data set names, standardized by a data set naming convention, are used to determine ownership. VTOC scans usually produce the necessary information.
VSAM data spaces, however, present a special charging problem. At sites that do not run IBM's Integrated Catalog Facility (ICF), VSAM naming conventions do not use the owner name in the data space name. Even when ICF is installed, the ICF Data in Index Component Name does not typically provide enough information for billing.
An additional complicating factor of identifying VSAM data set ownership is that a single VSAM data space may contain many clusters (data sets), which belong to separate users. Since VSAM catalogs contain unique entries showing the standard data set name for each VSAM cluster, the problem can be avoided by determining VSAM data set allocations by scanning the VSAM catalogs.
The information for non-VSAM data sets obtained by the VTOC scan and the information for VSAM data sets obtained from the VSAM catalog scans provide complete charging information for permanent DASD data sets. This combination of VTOC and catalog scans provides a standard approach for charging.
It is important to optimize the performance of both scans to minimize operational and overhead problems associated with gathering the data.
DASD charging uses space units such as volumes, cylinders, tracks, or bytes of data over a unit of time, such as days, to produce a charging unit of measure, such as track days (where one track day is equivalent to occupying one track for one day). The charging unit allows the DASD information to be extracted at varying intervals while keeping the same unit of reference. One representative unit of measure is the megabyte hour, which denotes the occupancy of one million bytes for one hour. The use of megabytes provides a consistent measure that does not vary with device type, unlike tracks, which are device-dependent; the timespan of hour provides more precision than day or month. While the traditional track day qualified by device type is still a viable unit of measure, it does not allow for as much flexibility as megabyte hour. The charging units multiplied by a rate per unit produce a charge for each data set, which is billed to the appropriate cost center.
Small data sets present a dilemma when the charge for them is so small (less than one cent) that it rounds to zero. Three alternatives to address this situation involve calculating the charges on a monthly basis, using a minimum data set charge, or carrying greater precision in the charge field.
Another important DASD charging consideration is that some permanent data sets may be used by many cost centers. The charges for these shared data sets can be distributed among multiple cost centers by prorating the cost. The installation sets the guidelines for determining how the charges are to be distributed.
As IBM's System Application Architecture evolves, it may be more reasonable to assign the cost of shared databases to a corporate account and view the data bases as a corporate asset. Another option may be to recover the cost with access charges to users.