From: Brian D Williams (talon57@well.com)
Date: Thu Jan 18 2001 - 12:20:28 MST
From: Brian Atkins <brian@posthuman.com>
>Brian you might want to watch this video:
>http://biz.yahoo.com/oo/010117/48967.html
>"Former GX CEO Predicts 'Crushing' Restructuring for Telecoms"
Thanks, I'll try to catch this.
>> No system being sold runs at anything near these numbers. I
>> recently got to play with OC-768 which is a mere 40 GBS. The big
>....which runs over a single wavelength of light among many many
>in a single fiber.
Known as a Lambda, I do this for a living....
>These technologies most certainly ARE commercially available.
>There are also terabit routers available to handle the packets
>(Charlotte's Web, etc.)
Excuse me but the industry standard is still just in the beginnings
of migration to OC-192.
Dense wavelength division multiplexing... simple.
I've gotten LIGHTWAVE since issue 1.
>> problem is that while 100 cables can theoretically carry the
>> traffic, the reality is that connections will need to be made to
>> many thousands of different points. We will need a thousand
>>times the Internet we have today. I also stated that these
>>networks will need to be built to accomodate QOS which they do
>>not now.
>Well this goes back to the old circuit vs. packet switching
>argument. I think you just throw more bandwidth at it, and QoS
>concerns go away. I mean if you have a 100mbit connection to your
>home, you really don't have to worry about making 64kbit phone
>calls. Do you really think QoS will matter in a world like
>that?
It's more than this, we don't have 100MBS to the home, that's the
point.
I don't think we'll ever have a 100MBS to the home standard, I
think we'll blow right by this. The problem is that the existing
network can't support this, and to build one today with existing
technology would run well over $1000 per port, times the 250
million ports in the 5 state region, we're talking quite a sum.
But you are right, eventually you'll pay to have a high speed port
and the amount of bandwidth needed for voice service will be so
small as to be negligible.
Something I've always thought interesting. As we advance to the all
optical network, we may start with a system that basically switches
individual fibers with MEMS, in effect replacing the copper with
fiber creating a mostly optical circuit switched network!
>
> As the bandwidth increases, eventually the cost of the bandwidth
> needed to conduct a voice call will be miniscule, but we are
> hundreds of billions of dollars from that point. TANSTAAFL.
>Exactly, it ain't cheap to get there. You might enjoy watching
>this video:
>http://biz.yahoo.com/oo/010117/48967.html
>"Former GX CEO Predicts 'Crushing' Restructuring for Telecoms"
>Translation: switch to IP and fiber or die
Thanks, got it the first time. ;)
>
> The days of saying the phone companies understand voice but they
> don't understand data are over.
>We'll see... T not looking so hot lately
Doing better lately. We are looking at all of this, but the numbers
don't line up, mostly due to the restrictions imposed by
regulation. If they hadn't of stopped us last time this would
already be built.
Brian
Member:
Extropy Institute, www.extropy.org
Adler Planetarium www.adlerplanetarium.org
Life Extension Foundation, www.lef.org
National Rifle Association, www.nra.org, 1.800.672.3888
Ameritech Data Center Chicago, IL, Local 134 I.B.E.W
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:05:00 MST