swithch and hub question

Taavi Burns taavi-LbuTpDkqzNzXI80/IeQp7B2eb7JE58TQ at public.gmane.org
Mon Jun 7 21:09:38 UTC 2004


On Mon, Jun 07, 2004 at 04:30:07PM -0400, Robert Brockway wrote:
> On Mon, 7 Jun 2004, Lennart Sorensen wrote:
> > the reason that they don't usually put them on computers, but I am
> > really not sure.  Maybe it's a cost thing.)
> 
> I wondered this too but I can't see how the logic could be very
> complicated so the cost should be a few millicents.

Keeping in mind that the logic has to ensure that nearly no noise is
added to the system.  It costs something in one-time design, but
probably nothing in production.

> could be customer demand or whatever.  Other examples of advances with
> this problem include cell phones that synchronise time off the network and

That kicks ass.  I was so happy when my celphone kept the correct local
time when I flew back to Alberta for a week. :)

> Anti-Lock braking systems in cars (which took decades to get to the market
> I understand).

Though so far as I know ABS is a BAD thing when trying to brake on icy surfaces.
The reasoning I recall is that grinding the tires along the ice with whatever
added sand is available actually _increases_ friction, whereas locking the
tires on water probably just promotes hydroplaning.

> One question about MDI/MDIX, what mechanism do they devices use to avoid a
> dead-lock, where both ends are trying to auto-sync?  Do they backoff and
> wait for a random amount of time before poling again, thus allowing for
> overlap windows in which the 2 ends can sync up?

I imagine it's something like that, given how collisions are handled (i.e.
the same way).

-- 
taa
/*eof*/
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml





More information about the Legacy mailing list