[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: DPD & DPV requirements



Skip,

Steve,

Judging from your response, I think I did both of us a disservice by
providing too many details from my stream of consciousness.  Here's the
distilled essence of what I intended to say:

In your requirements document you wrote, "Finally, this analysis suggests
that there are relatively few differences between a DPD and a DPV server ...
[T]he WG must decide if two distinct protocol formats, with different syntax
are warranted, or if a single [one] is appropriate."  You then went on to
conclude that they were sufficiently similar to justify a single set of
requirements.

Based on the way the discussions have gone, I'm questioning this conclusion.
By separating the two, DPV could focus on simplicity and efficiency (to
better serve highly constrained clients), while DPD focuses on completeness
(to serve the needs of highly capable, but perhaps contextually constrained
clients). DPV could consume DPD services, creating a complete solution to a
wide range of problems.

It is this last paragraph that I'd like comment on. Without some sense of
agreement at this conceptual level, it's fruitless to dig into details.


I think we are seeing some divergence in arguments for the two types of servers, wrt queries and responses. They may be small enough to use a unified format with options, or we may have two distinct formats with a lot of overlap. I'm not wedded to either approach. But, the need for both severs to perform path validation still seems to be true.


Steve