Yeah, I saw that...
With all respect to Dave, and not to sound too skeptical,
but we're pretty far along in our current architecture to
"fundamentally" change, don't you think (emphasis on
fundamentally)?
- ferg
'changed'. No need (and probably little likelihood now) to change
this architecture if you don't want to, but a new architecture may
come along that make this one seem quite outmoded.
I'm skeptical about something truly new coming from this specific
project, but I hope it comes from somewhere. With any luck someday
we'll be referred to as those 'old interphants'. 
John
It's also worth remembering that packet-switched networks took decades to eclipse circuit-based networks and that the early Internet was, for all intents and purposes, useless for the vast bulk of humanity in addition to being a tax-funded research project.
It takes a lot of seeds to grow a field of wheat.
Yeah, I saw that...
With all respect to Dave, and not to sound too skeptical,
but we're pretty far along in our current architecture to
"fundamentally" change, don't you think (emphasis on
fundamentally)?- ferg
Many people probably share similar concern.
My personal view (I've left MIT 16 years, so no relation to Clark):
- I believe we all wish the Internet architecture, as we have now, has
some problems here or there.
- But how to make it better? Quoting Dave, looking one incremental step
each time is unlikely the best way to proceed.
- To see see more clearly where we should head to, one can try
a 2-step approach:
+ if one gets all one's wishes: how would we want the architecture
to look like, given what we know today (that we didn't 30 years ago)?
+ if/once one gets that question answered, we can then tackle the next
question of how to get there from here.
my 2 cents,
Lixia
I'm skeptical about something truly new coming from this specific
project, but I hope it comes from somewhere.
the problem is that there are really no fundamentally new great
concepts. so this is likely doomed to be yet another second
system syndrome.
randy
Fergie (Paul Ferguson) wrote:
Yeah, I saw that...
With all respect to Dave, and not to sound too skeptical,
but we're pretty far along in our current architecture to
"fundamentally" change, don't you think (emphasis on
fundamentally)?
Most of the routing and security issues on todays IP4/IP6 internet could be solved by deploying HIP or derivatives thereof without requiring fundamental changes to the infrastructure since the major "flaw" of current generation Internet is tying the network identity and host/application indentity into one which is then overcome with whole spectrum of solutions along the lines of anycast, load-balancers, NAT, etc.
Pete
And the world demand for computers might someday approach 100?
How do we *know* there are no fundamentally new great concepts ...
unless we *try a lot of stuff*.
How many light bulbs did Edison throw away?
Cheers,
-- jra
How do we *know* there are no fundamentally new great concepts ...
unless we *try a lot of stuff*.
Trying stuff is good -- until something's tried, none of us can
really know what it'll do. At what point do entirely off-network
experiments become on-topic for nanog? (I doubt anyone has an
easy answer, I just wanted to throw the question out there.)
How many light bulbs did Edison throw away?
42?
To be fair to TJ Watson, please note that IBM was *already* engaged in
the production and sales of automated tabulating equipment, and when reading
his comment *in historical context*, it's pretty obvious that what he
*meant* by "computer" was "high end machine that only a few could afford".
In other words, what we now call a "supercomputer".
And sure enough, looking at the current Top500, http://www.top500.org/lists/plists.php?Y=2005&M=06
we see that only 6 sites have bought 20Tflops+ systems, but 19 are 10Tflop+,
and there's a *huge* pool of very similar smaller systems down in positions 300-500.
And this shape has remained remarkably consistent - anywhere from 3-7 systems that are
*way* out in the lead, a second string of several dozen smaller, and a huge pool
of lower-end machines. So TJ was totally right - at any given time, there's only
5-6 sites willing and able to buy that very top-end box....
That's atleast 2 orders of magnitude off:
http://www.ideafinder.com/history/inventions/story074.htm
interesting story though.
How do we *know* there are no fundamentally new great concepts ...
unless we *try a lot of stuff*.Trying stuff is good -- until something's tried, none of us can
really know what it'll do. At what point do entirely off-network
experiments become on-topic for nanog? (I doubt anyone has an
easy answer, I just wanted to throw the question out there.)How many light bulbs did Edison throw away?
edison didn't invent the light bulb...
So he didn't. And me a regular Wikipedian...</ot>
Cheers,
-- jra
But he DID make it more feasible and useful. And he DID throw thousands of
them away!
Scott