barns@GATEWAY.MITRE.ORG (06/05/90)
Some of you will recall a discussion a few months ago about OSI System Titles (initiated by a question from me). A short document was recently brought to my attention; I've typed it in below for your convenience (don't expect me to make a habit of this though). This sheds some light on the question of what a System Title looks like. ALSO, it seems that SC21/WG1 has someone working on a document called "PDTR xxxx - Information Technology - Open Systems Interconnection - Tutorial on Naming and Addressing". The copy I have was brought back from Seoul by our local traveler/victim, and only has a Seoul document number SC21/WG1 S 46. I suppose it will get a real number. This is a rather nice document which explains many things that I had to learn much more painfully. About system titles, it says: A system-title is used to identify a real open system. The system-title is a layer independent primitive name, i.e., it is used to identify a real open system as a whole. A single real open system is named by one and only one system-title. The system-title may be used in conjunction with other qualifiers to identify specific OSI resources in a real open system, i.e. it may be used as a basis to build structured names for objects in the given real open system - e.g. an application-process-title may be based on the system-title (with the addition of relevant qualifiers). And here is the other document: Date: March 1990 Title: Registration of system titles Source: UK NB contribution to SC21/WG6, Seoul, May 1990 The UK notes that work in SC21/WG4 on the structure of management information has demonstrated a need to make use of system titles in order to identify managed systems. This need is documented in an output document from the editing meeting on DP 10165-1 (in Paris, January 190) entitled "names for managed system managed objects"; this paper is to be allocated an SC21/WG4 reference number. Moreover, SC21/WG4 has identified a need for some means of system title registration. Consequently, the UK proposes that SC21/WG6 takes action to satisfy this registration requirement. As an initial contribution towards this work the UK makes the following suggestions: a) reliance should be placed on the general provisions of ISO 9834-1 regarding the composition and allocation of system titles, rather than developing a separate Part of ISO 9834; b) an amendment should be made to ISO 7498-3 (the Naming and Addressing Part of the OSI RM) that explains the approach to be taken for system title registration. Notes: 1. As implied by (a), two forms of system title name should be permitted, a Directory RDN nameform and an Object Identifier nameform. 2. Point (b) is intended to satisfy the requirement implied by reliance on the general provisions of ISO 9834-1 for registration purposes. ------------------------------------------------------------------------ When I consider the above document in light of the philosophy espoused in the Tutorial I referred to, it would seem that the RDN and OID forms of the System Title of a particular Real Open System have to be thought of as being one and the same System Title. This naturally leads to the question of which version ought to be used in the ASN.1 in various contexts, and whether both forms need to be understood by entities. For example, must an end system treat two A-ASSOCIATE-REQUESTs in the same way if one uses an application-process-title based on the RDN form and the other uses the corresponding OID form, or ... Seems that this is rather like the question about whether numeric and printable string organization names should be recognized as equivalent in X.400... /Bill Barns barns@gateway.mitre.org