Normally procurement understands by cost‐benefit analysis a comparison of the ‘total cost of ownership’ of an investment within a given time frame, for example 3 years, and the benefit achieved by the investment.
The cost (for software) is easy to calculate, it is the license and maintenance fees, plus schooling, training and usage (manpower cost).
The benefit consists of all kind of savings (if any) and other calculative achievements accomplished with the software. For BPA4DB2 one could calculate:
The savings in CPU time
This part is supported by the product. With BPA4DB2 it is possible to specify the company’s internal price of a CPU second. The tool calculates/prognoses with this and its measurement findings regarding the ‘achieved/achievable savings by tuning measures’ the amount saved per day. This pertains to the measured DB2 system and must be projected for all DB2 subsystems.
The savings in manpower
(a) BPA4DB2′s workstation and its ability to reveal tuning potential with one glance dispenses with database admins the need to reason and speculate about hit rate, page residency time, to try what‐if analysis, and what else competitor tools and particularly traditional monitors normally require. This part is difficult to calculate/estimate. Imagine a company with some hundred DB2 subsystems and a group of system DBAs versus a company having one DBA who is at the same time the MVS system programmer and the supporter of the SAP application.
(b) BPA4DB2′s Host Examination Component facilitates automatic observation of any number of DB2s. This feature can easily be the crucial point, the decisive item over the competitors. But the calculative benefit is again difficult to determine in general way. Imagine a company that did little in the area of bufferpools before the purchase of BPA4DB2 versus a company that sustains a group of DBAs to take care about IO performance. But in a given situation it is normally easy to compare investment and costing, it’s only difficult to present a general approach.
What about investment and costing for a fire extinguisher? A similar aspect is with BPA4DB2′s checks for ‘SCA fill level’ and ‘Group Bufferpool Write failures’. How is the calculative appraisal of a possible outage of an entire DB2 Data Sharing Group?