It seems appropriate to finish this paper with an account of some practical aspects of this project. A key resource for the development of an application intended to manage network nodes is (definitely) technical documentation. It is only fair to acknowledge Fore Systems for the extensive technical as well as tutorial documentation they make freely available at their web site. On the other hand, it was very difficult to obtain additional technical assistance through the channels set up with this purpose. One can't help but to ask himself if these channels are provided by the company with only customers in mind.
Anybody who has ever read MIBs knows that they are not always clear enough on their descriptions of tables and variables. Tests are often necessary in order to figure out the relationships between variables in different MIB tables, among other things. Given the fact that ATM backbone switches are hard to come by, it was a blissful event to find Fore's most advanced switch types at the Leibniz Supercomputing Center (LRZ).
While the people at the LRZ were very helpful, the workstation set apart for network management test purposes was running a Solaris 2.5 version, so that it has still not been possible to test the Technology Adapter on the LRZ's switches. An upgrade to Solaris 2.6 on the test machine --necessary to run the new version of the INMS-- should be completed any time soon.
Finally, an unexpected development in the ATM market took place at the time of writing this paper. Fore Systems announced in April that it has reached an agreement with General Electric to become a fully owned subsidiary of the electrical industry giant. The news that a direct competitor of Siemens takes over the vendor of the technology intended to be integrated in the INMS product line upon completion of this study project raises interesting questions. Whatever the outcome, this paper will not handle such non-technical matters.