Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 15 May 2001 18:30:34 -0700 (PDT)
From:      Richard Hodges <rh@matriplex.com>
To:        Jin Guojun <jin@george.lbl.gov>
Cc:        atm@FreeBSD.ORG
Subject:   Re: Marconi ForeRunner HE 155 and HE622
Message-ID:  <Pine.BSF.4.10.10105151815030.52241-100000@mail.matriplex.com>
In-Reply-To: <200105160111.f4G1Bq614194@portnoy.lbl.gov>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 15 May 2001, Jin Guojun wrote:

> A few years ago, the ATM seems the best one to build a WAN backbone.
> But the VC is painful and the overhead is high 5/48 ==> 10%.

Agreed.  It is quite an overhead when you have "only" an OC3 or less.
But if I had plenty of fiber, and OC48 or better it would probably 
not matter that much.  With all the fiber installed everywhere, why
are the phone companies not selling tons of bandwidth?  My guess is
that they don't have a market for the bandwidth, and would rather
sell off tiny slivers, one T1 here, one DS3 there...  I hear of so
much dark fiber, that it seems that the fiber owners are not really
concerned with the "cell tax" right now.

> The ATM will continue to server the middle bandwidth network, especially
> in the DSL and LAN area because it has existing mechanism to pace and
> divide the bandwidth. But it may not be good for high-speed backbone
> any more due to its high overhead.

When all the OC48 fibers are saturated, I'm sure efficiency will be
more important than today.  (Are they still bundling 4 OC48 lines
together to make an "OC192", or is OC192 finally a reality?)

> 
> Think if we put TV stations on network, put video rental stores on Internet,
> add visual telephone too, and we need a reliable service on the Internet.
> 
> Not just ATM QoS cannot provide enough control, IPv6 does not either.

I can set up an SVC with exactly 171 cells per second for a 64k voice
call.  Or a movie at exactly 23940 cells/second.  Or whatever I want...
And I know that the circuit will handle that traffic until I close the
socket.  Error rate is about 1 cell per 5 to 10 million.  That's good
enough for me :-)

> It will probably need bandwidth allocation mechanism in network layer,
> and put the network layer directly onto the WDM to get better performance.

By "it", you mean GigE?  IFF GigE ever does get quality of service
_guarantees_ (not just priority), and a point-to-multipoint facility
that works well with medium-speed video traffic (aprox 3 to 8 megabits),
then I will probably jump on the bandwagon :-)  But to get there, you
may have just re-invented a good portion of ATM.

All the best,

-Richard

-------------------------------------------
   Richard Hodges   | Matriplex, inc.
   Product Manager  | 769 Basque Way
  rh@matriplex.com  | Carson City, NV 89706
    775-886-6477    | www.matriplex.com 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-atm" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.10.10105151815030.52241-100000>