🚀 go-pugleaf

In article <qph7hls@sgi.sgi.com>, rpw3@rigden.wpd.sgi.com (Rob Warnock) writes:
>myoung@force.ssd.lmsc.lockheed.com writes:
>+---------------
>| I am still a little confused about concentrators and rings.
>| Is the general usage for FDDI to use a concentrator? Or do
>| users generally build physical rings?
>+---------------
>
>It varies wildly depending on the environement. Some users mix the two.
>
>+---------------
>| And of course the follow up:  If they use a concentrators, then why
>| pass tokens out to the stations, when the token can be passed about
>| inside the hub?
>+---------------
>
>Uh... Unless you pass the token to the stations, how're they gonna know
>when they can transmit?  That's what the tokens are *for*.
>
>
>-Rob
>
You caught me; I was asking a loaded question.

The token serves two purposes, which in a physical ring are combined:
1) The allow some implicit flow control, and
2) The control the distributed switch function.

In a hub switch, the two functions can be separated.  One can
add some buffering, and then let each line have its own asynchronous
token passed between the desktop and hub, signaling for data when
the buffer is near empty.

Meanwhile, the switch can have its own round robin token, taking
data from each hub interface buffer and switching it accordingly.

I wanted to see if any thread develops along this line of thinking.

Matt