T.R | Title | User | Personal Name | Date | Lines |
---|
1445.1 | | koning.lkg.dec.com::koning | Paul Koning, B-16504 | Fri Sep 16 1994 11:28 | 9 |
| Yes.
Full duplex is possible whenever there are exactly TWO MACs on the ring.
It doesn't matter what else there is; in particular it doesn't matter how
the ring is put together. So if you have a dual homed setup connecting
between a Gigaswitch line card and a DAS adapter, you have two MACs at all
times, and full duplex should work.
paul
|
1445.2 | retry: dual-homing and full-duplex | MUDDY::WATERS | | Mon Sep 19 1994 11:20 | 15 |
| > <<< Note 1445.1 by koning.lkg.dec.com::koning "Paul Koning, B-16504" >>>
>Yes.
>Full duplex is possible whenever there are exactly TWO MACs on the ring.
Another answer: Maybe.
You didn't specify what kind of GIGAswitch ports you are using.
If you connect one DAS host to two SAS GIGAswitch ports, you have
a 3-station ring with *three* MACs.
If you connect two DAS hosts to two DAS GIGAswitch ports, and swap
cables so that each DAS host talks to both of the DAS switch ports,
I'm not sure what you have. (An illegal topology?)
Paul, please explain in what case your answer applies.
|
1445.3 | DAS across GIGAswitch M ports? | CHEFS::PADDICK | Michael Paddick - BRS Bristol, UK | Thu May 22 1997 06:43 | 7 |
| If a DAS device is connected between two GIGAswitches with SAS ports
configured for type M, will the 'live' leg run as FDX?
The two GIGAswitches are inter-connected via another SAS port, so are
part of the same 'network'.
Meic.
|
1445.4 | ask the Gigaswitch folks | NETCAD::ROLKE | The FDDI Genome Project | Thu May 22 1997 10:20 | 13 |
| Hi,
Paul's answer in .1 applies. For any random DAS device which wants to
run Full Duplex the key metrics are that the "ring" has only two MACs
and that the peer whats to run Full Duplex, too.
Please see SCHOOL::GIGASWITCH note 491. This has a picture of what
you describe. I believe that Gigaswitch is very clever and can run
several Full Duplex links at once. How it does this is built in to
Gigaswitch software. I suggest that this question is best answered
there.
Chuck
|
1445.5 | | AMCFAC::RABAHY | dtn 471-5160, outside 1-810-347-5160 | Thu May 22 1997 10:23 | 23 |
| re .3:
>If a DAS device is connected between two GIGAswitches with SAS ports
>configured for type M, will the 'live' leg run as FDX?
Maybe. For FGL-4 ports in the right slots you will get full duplex. For FGL-2
ports in any slot you will get full duplex only if the other port of a pair is
not active.
M-AB-M the B-M link is active and can do full duplex, if the
B-M link fails the M-A link becomes active and it too
can do full duplex
S-AB-S is a wrapped ring and cannot do full duplex until one of
the S ports goes offline, naturally if the DAS goes down
the two S ports won't be able to see each other
A-B is a ring with two stations and will do full duplex,
B-A this remains true when a link fails and the ring wraps
A-BA-B is a ring with three stations and cannot do full duplex
B----A well, if a station leaves the ring it'll wrap and go
into full duplex
|
1445.6 | | NPSS::MDLYONS | Michael D. Lyons DTN 226-6943 | Thu May 22 1997 11:02 | 17 |
| ...as Dennis says, in the case where you make a GIGAswitch/FDDI DAS
port (FGL-2 configured with two PMDs in the port) into a port with two
M ports, you will get full duplex if only one MAC attached to those M
ports is active...
I.E. The MAC for a "A" port in standby is not visible to the M port, so
if you have an active "B" port connected to one of the M ports, and a
standby "A" port connected to the other, during normal operation FFDT
is operating to the "B" connected port. If the A port becomes active,
the port drops out of FFDT due to the presence of the third MAC.
SAS ports converted to M ports will always allow FFDT operation
since there is only one MAC at the other end. ...if, of course, the
other end allows it, and the GIGAswitch/FDDI port hasn't had FFDT
disabled.
MDL
|
1445.7 | | NPSS::MDLYONS | Michael D. Lyons DTN 226-6943 | Thu May 22 1997 11:05 | 4 |
| ...incidently, in the previous replies, .1 is correct, and .2 isn't,
for the reasons I mentioned in .6
MDL
|