Chapter 16: Choosing WBEM Software

If you are about to embark on a WBEM/CIM project within your company then you will obviously need to select the appropriate WBEM server software. This raises the normal "build or buy" debate and it is perhaps worth looking at the effort involved in building and supporting the basic WBEM software.

Home Brew

As the standards are firm you may be tempted to build your own WBEM server and associated tools. For your particular products this may be the best way to go but, when estimating the size of the development, bear the following numbers in mind.

The openPegasus release 2.2 code consists of about 114,500 non-comment lines of C++ broken down as shown in Table 16.1. Although this represents only one implementation, I have no reason to believe that it is atypical.

Table 16.1: openPegasus v2.2 Lines of Code

Module	Approx Lines
Server	102,000
mof Compiler	6,500
Client Libraries	6,000
Total	114,500

If the COCOMO II ^[1] model for estimating the time required to develop software is retrospectively applied to this, and if appropriate values are taken for the productivity factor (2.58) and the penalty (1.11), then this represents an effort of about 40 programmer years .

Continuing with the COCOMO II model, the effort required to maintain this amount of code can be estimated at 14 programmers in the first year, dropping to a steady-state maintenance team of 7 programmers after 4 years.

These are fairly large numbers for any company and, if you are thinking of developing your own WBEM system, you need to be willing to spend this amount of money. On the other hand, if you buy commercially, then you need to ensure that the supplier has dedicated a team of approximately this size to the ongoing maintenance of the product. Remember that these numbers are for the WBEM server only; they do not include any allowance for writing your own application ”the clients , providers and listeners.

^[1] See http://www.softstarsystems.com/cocomo2.htm.