Eric, I'm in a state of wonderment over how you mis-
interpreted what I intended to convey. I used the original
Apple and Wang models in order to characterize closed
architectures, not open ones.
My reference to *closure* (which I bolded in my original
post citing the fool's references) was a double entendre, with
what I thought were obvious tongue-in-cheek overtones. I
guess there is a legitimate use for those emoticons, after all.
But while we're on the subject, the use of Apple does serve
some better purpose, and that is that Apple has realized,
after experiencing some extreme pain, the evils of their
former ways. Today's Apple is more open, by any stretch of
meaning of the term "open," than they were during the period
in time which the fool was referring to. Their reference, in
other words, was historical in nature, as was mine to Wang.
But you've opened the door to a good discussion,
nonetheless:
What does open really mean, anyway? And in how
many different strata or dimensions, how many contexts,
should we be examining it?
In the discussion at hand, open has been used in two distinct
contexts, at least:
(i) the open which relates to business policies of
admission and denial controls pertinent to the other
ISPs attempting to gaining access to the MSOs' networks,
and then there is,
(ii) the open which refers to the application interfaces being
used at all layers of networking and their underlying
purposes, which should be in support of human and
machine communications, in all possible combinations.
Beyond the physical interfacing, there are open and shut
issues at every layer of the OSI Reference Model and
TCP/IP stacks.
The first of these (i) [admission and denial controls] is the
one which I think is currently being addressed both here and
in the popular press, as it speaks to the admission or denial
of multiple ISPs and other players to the MSOs' networks.
But I think that the second context (ii) [conformity to stacks
and services] must also bear more than just a modicum of
mentioning, as well, for as the former results in major
increasing disparities, the consequences to the latter will be
felt in kind.
-----
Yes, DOCSIS is an open standard, but within the specific
scope or dimension that is separate from those which affect
end user protocols and ultimate communications. DOCSIS
is the manner in which the cable industry has elected to
manage physical media transport and termination, in
harmony with their cable modem termination system
(CMTS) standards at the other end of the wire, the head
end.
But DOCSIS standards do not directly address the
standards which are used to support end-user interplay with
one another, or with the WWW at large, per se. At least
they shouldn't be, despite the manner in which they
sometimes intervene.
DOCSIS is not a standard, in other words, in the same
context in which ratified RFCs are, which are the IETF
protocols and practices that must ride and partake over the
MSOs' facilities, in ways that are supposed to be consistent
with how they traverse all other forms of access and
transport, through other venues.
These standards and rules from the IETF and other bodies
take place at different layers of the stack. In some cases the
DOCSIS rules compete or are otherwise dissonant with
those of the Internet's, at the same time. And there's where
they'll get in trouble, eventually, if they find themselves
departing too independently, or too radically.
If and when MSOs' networks don't align with the protocols
that must traverse them, you wind up with something that is
less than a homogeneously standard architecture, but that is
okay since it plays to the theme, and is consistent to some
extent, with other "network of networks" principles. To a
point.
But when the discontinuities become extreme, demanding of
a separate set of server paradigms in order to mend radical
differences, those which ascend from different philosophical
roots (as I think is quite plausible if some of the proprietary
folks don't begin aligning, soon), then there is sure to be
trouble for them and their end users, down the road. I say
this because synchronizing changes between disparate
architectures can be a horrible bear to wrestle, which often
results in a 'no contest' kind of withdrawal from the match..
It's more like a closed architecture at some point, when
special provisions like proprietary gateways are built in,
which are required in order to translate intranet primitives to
those of the rest of the world's Internet.
Here I'm referring to intranet border gateways, specialized
DNS servers, proprietary policy enforcement, non-standard
forms of caching, and other mostly administrative functions,
although a growing number of transport layer disparities may
arise, as well.
As an example, consider Open Packet's posture for
provisioning voice services over cable, against the rest of
what's going on in that space directed at IP Telephony
initiatives on the greater public 'net (and, oy! lest I forget,
the PSTN's modified VoIP initiatives, as well).
You say,
"Another major issue that people keep ignoring is that
opening the wires to multiple ISP's does nothing to change
the monopoly control of the MSO's in how they build their
systems and who they buy equipment from."
I see. You're saying that, If an MSO decides that they
don't want to be forced into meeting oncoming demands in
the same ways in which the rest of the world is moving, then
no one can force them to. [I find myself striving to come
up with a Yogi Berraism at this point to match that kind of
logic. Anyone?]
That's how I interpret what you have stated, although I think
that you are merely reflecting the realities as you see them
existing today, which I would agree with, and not your own
feelings on the subject. Correct me if I'm wrong.
In any event, my take on your statement, which I've already
stated above, is that it is consistent with the ways in which
monopolies have treated such issues going back to forever.
-------
Again, I did not use Apple in order to point to an example of
openness. Rather, I referred to them in an historical context
to point out just the opposite.
Would anyone else like to offer a rendering of what they
think "open" means?
Regards, Frank Coluccio |
