At 04:51 PM 9/10/2001, Eamon O'Tuathail wrote:
>> Frankly, I don't see what the "abstraction of the transport
>> layer" provides you.
It means you don't have to define the same things again and again.
We should not be pursuing a debate about the utility of BEEP. That
question was settled already.
Folks who do not understand the issues are, as you very considerately note,
able to peruse quite a bit of educational material on the topic. However
the topic is irrelevant to this Last Call.
At 05:38 PM 9/10/2001, Christian Huitema wrote:
> Christian, perhaps you are forgetting that having multiple, related data
> streams use multiple, independent TCP connections is a problem, not a
> feature? Hence there needs to be multiplexing on top of TCP.
Either they are multiple streams or they are not. If they are not,
multiplexing them on the same transport connection is a bug, not a
feature.
oh. you want to ignore the problems that multiple, related tcp connections
cause? I thought the seriousness of that problem was established 8 or so
years ago.
again, you are attempting to make basic attacks on beep, but beep is
already standards track.
> And your knowledge about fundamental networking rules has demonstrated
> a significant disparity from the knowledge of quite a few other
> folks. Any literature references to this fundamental rule would
> be appreciated.
The effects of layered multiplexing were mostly experienced when trying
to run stuff TCP/IP on variation of protocols doing their own flow
control
Interesting. CSNet (the predecessor to NSFNet) ran TCP over X.25 just
fine. (There are lots of other examples, but that one rather stands out.)
However, yes, layers of flow control are an issue.
And BEEP deals with them, very explicitly and very carefully.
Do you have evidence that the beep specification does not work? If so,
please do bring that information to the beep working group. They will be
delighted to receive the feedback.
Again, however, that is not an issue for this Last Call. That is why I
have changed the Subject line, so no one will mistake this exchange as
being relevant to the Last Call.
D. L. Tennenhouse, "Layered Multiplexing Considered Harmful," in
Protocols for High-Speed Networks, Rudin and Williamson (Ed.), North
Holland, Amsterdam, 1989.
I love the title.
It suggests that TCP over IP over Ethernet is harmful, since each one does
multiplexing.
A similar problem leads us to avoid IP layer fragmentation
, because it
interferes with TCP layer reassembly and acknowledgement
The only problem with IP layer fragmentation is efficiency and holding
receiving host buffers for a long time. That efficiency hit is
substantial, but the functionality works fine.
d/
----------
Dave Crocker <mailto:dcrocker(_at_)brandenburg(_dot_)com>
Brandenburg InternetWorking <http://www.brandenburg.com>
tel +1.408.246.8253; fax +1.408.273.6464