SONET Interconnect (was RE: MCI)

Mike Trest trest at
Fri Mar 29 17:20:00 UTC 1996

The discussions regarding ATM/SONET and IP over ATM are finally focused
on a fundamental issue:

> Unless you begin building massive [native] long-haul ATM networks, this
> is not an acceptable transport for the reasons I mentioned earlier.
> - paul

IMHO, "... massive [native] long-haul ATM networks..."  are aready 
being put together (as we are doing)  because:

(1) ATM is based on "good science" - a lot of people did  a 
    lot of good research before building the first ATM switch

(2) ATM is based on "good economics" - pay large $$ once to
    put fiber in the ground. Pay smaller $$ increments for 
    speed improvements whenever you figure how to build better 
    electronics at ether end of the glass.

(3) ATM is being supported by "good people" - I am not surprised
    by the hundreds of announcements comming from a variety of
    vendors.  Can that many manufacturers be wrong?  [I admit 
    that a few are quite naive regarding impact of doing IP 
    networks over ATM today.]

(4) ATM and IP are already enjoying "good success" together.  
    We all hear about problems.  Not enough is heard about 
    success. [Not enough press sizzle].  I have encouraged
    all our vendors and customers to put their ATM success 
    story on their WWW page.   Guess what! Most are so busy
    with new business that they do not want to let their
    competition know how they are doing it.

Here at ATMNET we are biased in favor of using ATM for transport 
of IP traffic as well as non-IP traffic.  We are not stopping 
at the use of ATM as our backbone technology.  We provide ATM 
all the way to the customer premises at OC3 (or higher) speeds.  

IMHO,  ATM and IP are *NOT* conflicting technologies.  The common 
goal is to produce a survivable, scaleable, robust networks.  
The respective focus is on a different piece of the common problem.
Early ARPAnet researchers and TCP/IP network builders understood 
that the protocol had to be independent of the host computers and 
the transport mechanisms to endure.  [Yes, I date myself with this
admission.]  Today's commercial developers still understand
this natural division.  As network operators we must contribute
to the continuing IP/ATM dialog and focus our unique perspectives
[i.e. operating profitably] on the common goal.

I think members of NANOG are correctly voicing concerns about 
the future of both IP and ATM.  BUT have you noticed that a lot 
of ATM FORUM members are also the very same manufacturers who 
provide us our IP based equipment?  

Perhaps we all need to do a better job of telling our IP equipment 
providers of our ATM concerns.  This is my practice.  Unfortunately,
the manufacturers could all improve their responsiveness to
today's IP routing probelms.  

We try very hard to keep our vendors informed of our expectations 
of them on ATM and IP matters.  This way, they go the the ATM 
working groups and fight for what *WE* want.


Mike Trest                  EMAIL:    trest at
ATMNET                      Voice:    619 643-1800 / 619 643-1805
5440 Morehouse Dr, #3700    Fax:      619 643-1801
San Diego, CA  92121        Ans/Page: 619 960 9070

More information about the NANOG mailing list