> It's possible they might get there.  Engineers are pretty creative people
> about getting
> around the perceived limits of reality.

That's kind of how I'd like to interpret it: that while multi-mode
will always be cheaper for short runs (due to the ability to use
dimmer transceivers) and slower for long tuns (due to the multi-modal
dispersion problem) than single-mode. *But* maybe that won't matter so
much, since we tend to be comfortable using "one-notch below the
maximum" speed at any given time.

> So one thing you can do is to run a cable with multiple multi-mode fibers.
> Light one
> strand on gigabit, then if that link starts getting saturated, light a
> second and bond
> the links at the switch.

Already planning on this.

Rather than trying to have a single high-speed link between two "core
switches"—one at each campus—and distributing bandwidth out from
there, I intend to have multiple switches (2 to start, then more) at
campus B Gigabit linked to a "core switch" at campus A which will have
several, bonded, copper Gigabit links to our server switch.

Thanks to both Andrew and Rob for your thoughts!

_______________________________________________
Fwlug mailing list
[email protected]
http://mail.fortwaynelug.org/mailman/listinfo/fwlug_fortwaynelug.org

This is a public list and all posts are archived publicly. Please keep this in 
mind before posting.

Reply via email to