Well written post over at Raging Bull.com
ragingbull.com
Tellabs systems perform circuit switching. A circuit uses bandwidth *all* the time, whether it is
in use or not. ATM, Frame Relay, IP and other packet/cell/frame based technologies allow you
to take massive advantage of statistical multiplexing.
EG: If I have a certain amount of bandwidth, lets say OC-12c (~655 Megabits/s), then I can
circuit switch components of that bandwidth to another circuit (say, peel off a T3-45Mb/s), but
once thats done, the bandwidth is allocated. I can peel off 12 T3's and then I can't use that
OC-12c for anything more. But, if I'm taking advantage of packets/cells then I can oversell
that capacity. In other words, I can have 100 T3's (or however many I want) coming into an
ATM switch that only has an OC-12c upstream or "trunk" connection. Since I can bet that
not everyone will be using all their capacity at the same time, this works- within reason.
This is what some people call the magic of packet-switched networks.
Problems arise when you actually want to provide some real quality of service through such
a network. All kinds of issues begin to evolve that simply were not problems when you only
had circuit switched networks.
I'm not sure what you mean by an all-optical/digital networking environment. As regards taking
advantage of DWDM and such, Ascend is definetely very well positioned. Tellabs and Ciena
are a pretty good match. Unfortunately, there likely will never really be such a thing as an
optical switch. What people talk about as wavelengths, channels and windows on DWDM
gear today is in reality SONET frames. Yes, its very different from a good 'ol Nortel, Lucent
or NEC SONET system, but its still SONET frames.
I think the real question you may be asking is- how good are Ascend's switching products, and
how well will they do in the marketplace of the future. As to how well they do, that certainly
depends on a lot more than how good the products are. As to how good they are?
The GX-550 is the best you can buy today. Ascend really understands the *real* paradigm
shift, which is towards DWDM and IP. Believe it or not, and most people will laugh at me,
the 550 may well be a better *IP* router than Cisco's current state of the art.
Will it get better? Cascade and Ascend both had good reputations as innovators, and have
continued that after the merger; so I'm betting they will. Of course they will be leapfrogged
by their competitors, but then they will take their turn on top again. The real issue is how
well will Ascend manage to handle the product developments, acquisition and marketing.
And in that area, they're facing the best in the business. Cisco is frequently reffered to
externally and internally as the best technology acquisition and marketing company that
has ever existed; thats something that Ascend could learn a lesson from.
Now, whats this rumor I've heard about LU and Monday AM. Cruel games? :^)
(Voluntary Disclosure: Position- long; ST Rating- buy; LT Rating- buy)
|
