Re: Errors on Token Ring line

[Larry Bolhuis <lbolhui@xxxxxxx>]

Doug,

> As a quick test though, I dropped it down to just two
> devices on the ring.  The change was insignificant, about 0.05K/sec.

  And that's probably actually within the variance you would experience between
transfers!

> Guess which they like better? <g>

  How many guesses do I get???

> Besides, I'm not even sure a 100M card is available for their
> CISC machine.

  Correct, It's not, nor will ever be!

> My 400 adapter does support full duplex, as do my PC adapters.  But I
> have it set to half duplex, because my understanding was that to use
> full duplex you needed just two devices on the ring or a switch like
> you mentioned.

  You are correct.  Full duplex tells the adaptor that it can freely transmit at
any time because the other (there can be only one) device is always ready to
recieve.  Thats where the speed comes from, you never need to wait for the 
token,
just start sending even if you are recieving at the moment. Its really not a 
Ring
anymore at that point but who cares!

> If full duplex helps eliminate line turn-around time, then maybe it
> would help.  Maybe I'll try some experiments someday, but my gut
> feeling is that in my environment it won't buy me much.

  Not much indeed!  If data is primarily going one way (at any given time) then
full duplex won't by much more than a couple percent at best.

-- 
Larry Bolhuis         |         What if it's True?
Arbor Solutions, Inc  |
(616) 451-2500        |  
(616) 451-2571 -fax   | 
lbolhui@ibm.net       |
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---