Cisco HFR

Tue May 25 19:01:03 UTC 2004

> Date: Tue, 25 May 2004 10:18:39 -0700
> From: "Michel Py" <michel at arneill-py.sacramento.ca.us>
> 
> > This box is the HFR which, according to the San Jose Mercury,
> > is short for "Huge Fast Router". (Some reporter at the Merc
> > probably still believes in the tooth fairy.)
> 
> Same as the BFR, I heard a different interpretation of the word in the
> middle :-)
> 
> 
> > As with many things, if you have to ask how much it costs
> > before deciding to order it, you can't afford it.
> 
> You don't get it, me thinks. For lots of people here, networks are not a
> toy funded by the taxpayer's money, they're a tool to make money and
> popular wisdom that I have found being practiced here says that indeed
> you _do_ ask how much it costs before you buy it. I don't buy a million
> bucks peace of equipment because it looks cool and I just got funding. I
> buy a million bucks piece of equipment because I want ROI on it, and if
> the ROI is 500k over 5 years I actually don't buy it.

No, you don't get it (though I'll admit the statement is hyperbolic). 

If you need OC-768 connectivity, there is only one place to go today. If
you really need it (and few do), you have the money to pay for it and a
business that does not make existing hardware appropriate.

I tend to suspect that the need for OC-768c is still rather limited.
Most providers are primarily aggregators of massive numbers of small
pipes. With almost no OC-768c Sonet gear in the field, my guess is that
this gear will be used mostly by a handful of carriers who can make
fiber available at minimal cost for dark fiber connectivity at major
aggregation points.

In the vast majority of the places where more bandwidth is needed, just
adding more OC-192s (or even OC-48s) is often a more economical answer,
especially if you own lots of excess fiber and/or WDM gear.

As the the comments about "toys", yes, our network is taxpayer funded.
It is run for the USDOE with federal funding; no CA state funds are
involved. It is NOT a play network. It is a full-production high
availability network which provides the only access to a number of very
large and, in the view of some, very important facilities. It is an
unusual network in terms of the volume of point to point flows. Things
like fusion and high-energy physics generate simply astonishing amounts
of data. Terabytes are not uncommon. We don't need OC-768 today, but we
expect to need it in the next few years.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634