Cell Relay Archive

[Jerome BERNIER <BERNIER@frcpn11.in2p3.fr>]

On Thu, 6 Apr 1995 12:36:18 -0500, Thu, 6 Apr 1995 17:21:25 GMT COASTCOM said:
>I have been monitoring this UseNet for some time now and have not seen
>many articles posted indicating how many folks out there actually plan to
>use ATM and what specific problems ATM solves for them that cant be
>solved by existing technology.
>
>>From the vendor community, I have seen future applications that could be
>utilized, once the services are available from carriers and the
>applications software is available from the publishers.
>
>>From the user community, it seems for the most part that the early
>adopters are using ATM to transport huge amounts of data for applications
>such as seismic data and real time computing between super computers.
>This is fine for the companies that have these applications, but these
>applications arent really what you might call main stream.
>
>This again leads me to ask who plans to use ATM? I would appreciate any
>inputs on the following:
>
>1. Do you plan to use ATM?

We use ATM.
We use it on LAN workstations farms (cpu servers, disk servers,
tape servers) for intensive computing and data analysis
(High Energy Physics researches).

>
>2. What problem does ATM solve that current technology does not?

Switched high throughput.
(switched FDDI would be another answer,
but for us the ATM solution was cheaper and more promising)
By the way we think we will need 622Mb/s interfaces next year...

>
>3. What interfaces do you plan to use toward the public network? T1, T3,
>OC-3?

Today, only LAN needs.

>
>4. What kind of aggregate data rate is required on the ATM? 100 Mbs - 1
>Gbs? Greater than 1 Gbs?

Today ~20 workstations with a 100 or 155 Mb/s interface
---> theorically 2 Gb/s, practically 100-150 Mb/s

Next year --> x2 ?

>
>If you would prefer to send me a private e-mail, as opposed to posting an
>article, that would be fine. I have no preconceived notions on this, all
>Im trying to understand is if ATM is a mainstream technology for end
>users, or whether it something that will be prevalent among carriers and
>a niche technology with end users.
>
>Doug Stewart @ Coastcom
>