T.R | Title | User | Personal Name | Date | Lines |
---|
1215.1 | Yes, that configuration is supported | NACAD2::CORTES | | Wed Jul 13 1994 11:50 | 21 |
| Burance,
That configuration is certainly supported by by DEChub 900MS.
As a matter of facts, you are only using 12 out of the possible 14
flexible cannels available for this. Keep in mind, however, that
history of insertion, and slot position is important. In other
words, if you start from scratch and connect all your
concentrators from left to right, each pair being in contiguous
slots, you should see no problem. On the other hand, if previous
to doing this configuration you have done several LAN hopping
changes, and your concentrators are connected so that each pair
do not reside in contiguous slots, then you might see the situation
where connections will be rejected. This could happen because you
might end up with only split channels available when your connection
requires full-length channels. Remember that out of the 14 channels,
4 are split channels, i.e., they only run through half of the slots,
2 running from slot 1 to slot 4 and 2 running from slot 5 to slot 8.
Eladio
|
1215.2 | how about realistic | HGOVC::BURANCELEUNG | | Wed Jul 13 1994 22:47 | 15 |
| Hi Eladio,
In a network seminar, I heard that only 6 flexible ethernet and
2 FDDI flexible channels is supported for the time being.
It is proved that no more than 6 flexible ethernet channels
can be created.
Since I don't have enough equipment to test the maximum number of
FDDI channels, are you sure that under the current technology
(not theoretical) 12 FDDI channels can be supported ?
Thanks,
Burance.
|
1215.3 | These are facts, not theory | KALI::CORTES | | Thu Jul 14 1994 11:46 | 22 |
| Burance,
I am not sure why they told you that in the seminar. Right now all the
14 flexible channels in the backplane are available for FDDI. It is
true that only 6 channels are available for Ethernet, and I believe it
is done that way because of a repeater limitation rather than a hub
limitation. I am conducting system test for the DEChub 900MS and for
the DEF6X concentrator and believe me, I have created similar
configurations many times. As a matter of facts, I have had
configurations where all 14 channels are being utilized. This can be done
with a hub fully populated with 8 concentrators. Since 8 concentrators
configured Trunk AB will require 16 channels, the only way of having
all of them in the hub is by configuring 6 of them Trunk AB, one Trunk A,
and one Trunk B. I order to complete the ring you need to externally
connect the Trunk A and Trunk B concentrators. However, all the 14
internal channels are being used, and it works pretty good! This is not
theory. It is a real configuration that we have tested many times in the
lab.
Eladio
|
1215.4 | One "leg" = one channel | HGOVC::BURANCELEUNG | | Fri Jul 15 1994 08:47 | 23 |
| Hi Eladio,
Thanks for your solid answer.
>This can be done with a hub fully populated with 8 concentrators.
>Since 8 concentrators configured Trunk AB will require 16 channels,
>the only way of having all of them in the hub is by configuring 6
>of them Trunk AB, one Trunk A, and one Trunk B.
From your description, when DC900MX is configurated with AB pointed
to the backplane and occupies 2 channels, should we see 2
"red leg" attached to the DC900MX icon ?
I have tried that there is only one "red leg" labeled with "1"
attached to the icon. But does it mean that it occupies
2 channels ?
If there are 2 concentrators with both AB pointed to backplane,
how can we connect A-B and A-B (2 seperate channels) ?
Thanks,
Burance.
|
1215.5 | Firmware version | HGOVC::BURANCELEUNG | | Fri Jul 15 1994 08:48 | 4 |
| supplement to .-1
I use HUBwatch V3.0 and DEF6X200.bin for DC900MX.
|
1215.6 | One Leg Does It All | LEVERS::DRAGON | | Fri Jul 15 1994 09:09 | 16 |
|
Burance,
Regardless of whether you have A, B, or AB to the backplane
HUBwatch will show you the one "red leg". So it has 2 states,
no "red leg" or one "red leg". You connect the module to the
FDDI LAN by pulling that single leg down to the LAN. HUBwatch
and the MAM do the rest. I think that having 4 seperate icon
states might be clearer (one of which is no leg for AB out the front).
It would help someone looking at the LAN Interconnect Window
know whether their A or B is going out the front, without needing
to pull up additional windows.
Bob
|
1215.7 | MAM does it for you.... | KALVIN::CORTES | | Tue Jul 19 1994 16:58 | 18 |
| Burance,
As Bob says, Hubwatch will always show only one "leg" regardless
of your configuration. To know if it is trunk A, B or AB you have to go
to the Station Configuration window and look there. I certainly believe
that it would be much better if we could tell the configuration directly
from the LAN Interconnect window.
If you have two concentrators configured Trunk AB, and connect both
to the same LAN (ring), the MAM will do all the work for you, i.e., you
will end up with a correctly connected ring which, by the way, will
consume 4 flexible channels.
Regards,
Eladio
|
1215.8 | 4 ?? | PADNOM::PEYRACHE | Jean-Yves Peyrache Country Support Group France | Wed Jul 20 1994 11:50 | 6 |
| Elodio,
can you explain (may be with a little picture) why 4 IMB are reserved
for this particular config .
thanks
|
1215.9 | 2 Channels for each FDDI Connection. | NACAD::CARRAFIELLO | | Wed Jul 20 1994 17:14 | 13 |
| Each FDDI backplane connection requires 2 flex channels, one for
transmit data and one for receive data. The FDDI signal definition for
a flex channel is 5 bits of data and 1 bit for clock, thus using all 6
bits of a flex channel. The FDDI channels are unidirectional, therefore
one flex channel is needed for the transmit direction and one flex
channel for the receive direction.
The previous note mentioned a Trunk AB connection. This is a DAS
connection. Thus each connection (A to B and B to A) requires 2 flex
channels. Therefore the full Trunk AB connection between two modules
requires 4 flex channels. A SAS connection (Trunk A or Trunk B) would
have only required 2 flex channels, one for transmit and one for
receive between the 2 modules.
|
1215.10 | 16>14 | CLPR01::PEYRACHE | Jean-Yves Peyrache Country Support Group France | Thu Jul 21 1994 05:06 | 5 |
|
as you say, if i put 8 concentrators in DAS i consume 16 IMB
in HUB900, as this configuation i'snt supported ???
Jean-Yves
|
1215.11 | Won't configure more. | CGOS01::DMARLOWE | Have you been HUBbed lately? | Thu Jul 21 1994 11:35 | 4 |
| I think 14 FDDI is the max so something should break if you go beyond
that.
dave
|
1215.12 | only seven concentrators supported ,,, | PADNOM::PEYRACHE | Jean-Yves Peyrache Country Support Group France | Thu Jul 21 1994 13:59 | 8 |
| Dave
>>> I think 14 FDDI is the max so something should break if you go beyond
>>> that.
as you say 7 concentrators is a upper limit in DAS ???
Jean-Yves
|
1215.13 | 8 FDDI DAS = wrapped ring | NACAD::PARISEAU | Luc Pariseau | Fri Jul 22 1994 10:03 | 9 |
|
You can put 8 concentrators in but you will get a wrapped ring unless
you connect to the "outside".
But most people will setup one concentrator as a Trunk_A, 6 Trunk_ABs,
and one Trunk_B. And then you take the whole HUB and connect that
to the trunk ring in the building.
Luc
|
1215.14 | ok for this map | PADNOM::PEYRACHE | Jean-Yves Peyrache Country Support Group France | Fri Jul 22 1994 12:50 | 6 |
|
Thanks luc,
i agree that's works in this configuration
Jean-Yves
|
1215.15 | Does this mean I win something? | BUDDIE::KORNS | | Thu Aug 18 1994 00:08 | 8 |
| Gee, until I entered this reply and screwed up the number of
replies count, the "Replies" column had the correct answer
to the question asked in .0's title.
Never mind.
Dave
|