tebbutt@RHINO.NCSL.NIST.GOV (John Tebbutt) (01/18/90)
Dr. Rose, First, many thanks for your explanation of interim network address encodings. Second : >Thus, if "interim network addresses" are in use, you can >algorithmically determine which community is being referenced... Does ISODE do this ? In particular, can I, as an application process, treat an interim encoded network address *exactly* as if it were a legitimate OSI NSAP address, and load the ISODE struct NSAPaddr accordingly ? Does ISODE check the value of the OSI NSAP to determine what kind of network service is required ? Or does the user need to parse the interim encoded address in order to properly load the struct NSAPaddr of the Presentation Address structure ? If the latter is the case, does ISODE provide routines for parsing and encoding the interim form ? Third : On reading the ISODE 5.0 Manual, I thought for a moment that the naddress "normalization" routine, na2norm(), would provide the answer to my second point (above). Having read it, I am not so sure (!). What kind of naddress should na2norm() be applied to, and what kind of naddress does it return ? The implication is that it should be applied to addresses containing interim address encodings (presumably carried as regular OSI NSAPs), to convert them to `"real" OSI addresses' - I thought the interim encoded addresses WERE the "real" OSI addresses ! Thanks in anticipation, John Tebbutt, NIST
mrose@CHEETAH.NYSER.NET (Marshall Rose) (01/18/90)
The algorithmic interpretation of interim network addresses is fully implemented in ISODE 6.0. Basically, if you look at Kille's format, you notice that there is a community name field that is a number. In 5.0, this was the only format understood and the number mapped to a TS-stack. Further, in 6.0, the ISODE lets you define your own formats (within certain limits), and as a part of that you define the community name and the TS-stack to use. With this, you now have a fairly generic approach. For example, if you have an isolated TCP/IP LAN, you can easily define an OSI community for it. As an application process, you are supposed to treat network addresses as opaque. I suggest you read the 6.0 documentation set (when 6.0 is out next week) to get the answers to the rest of your questions. /mtr