dbh at enterasys.com
Mon Jan 6 11:35:51 CET 2003
I have suggested a number of times that we need better architectural guidelines.
Let me recommend that each Area should have a working group that develops a document identifying the key architectural elements that should generally be followed by the working groups in that area. For example, for the area I am most involved in, the O&M Area, it would very helpful to have a general architectural understanding of how various network management protocol efforts should fit together, including SNMP, COPS/PR and XML for device configuration purposes. The Applications Area should have a document describing how XML should be used within Internet efforts (and does have such a document) and one for LDAP. The Security Area should have one for security concerns. and so on.
Note how there is some correlation between where directorates are needed and my suggestions for architectural guidelines? I believe documenting the "common wisdom" of the areas would be helpful.
Such architectural guidelines would be useful for efforts in other areas to understand how to design their protocols to meet the guidelines set up by other areas, to help improve integration between protocols developed in different areas. In the O&M area in particular, it might be useful to help guide the integration of the Operations Area and the Network Management Area, since we seem to have never actually integrated the two areas into one.
I have recently been involved in numerous efforts outside the O&M Area that focus on configuring protocols on devices, and find that often the WGs have little understanding of the architectural guidelines fairly well understood within the O&M Area, such as the security requirements for network/device management (which were discussed extensively in a Security Area BOF but were never actually documented, to my knowledge).
I believe some of the IETF areas' "common wisdom" should be documented as architectural guidelines.
> -----Original Message-----
> From: Natale, Robert C (Bob) [mailto:bnatale at lucent.com]
> Sent: Monday, January 06, 2003 11:16 AM
> To: Margaret Wasserman
> Cc: problem-statement at alvestrand.no
> Hi Margaret,
> > But, I don't believe that we can break every complex
> > problem into a set of small problems.
> I agree entirely with the intent of that statement.
> However, the literal fact is that we *can* (not that
> we "should") decompose a complex problem into a set
> of constituent problems. Coupled with the natural
> tendency of the general make-up of the IETF toward
> taking that kind of "analytic" approach, that is more
> often than not the course that is taken. And then,
> very often, one or more of those "smaller" problems
> loom larger than before and the process is repeated
> again. It's not always obvious and it's not always
> problematic...but it is surprising how much WG
> effort/output in the end seems focused on such
> "smaller" issues (when compared to the "larger"
> original overall mission).
> > And, there are some complex problems, such as the
> > scalability and security of the Internet that I don't
> > think we should ignore.
> I agree...and those kinds of problems are where
> increased proactive leadership (from the IAB, IESG,
> and WG chairs) is needed, IMHO.
More information about the Problem-statement