| Title: | DEChub/HUBwatch/PROBEwatch CONFERENCE |
| Notice: | Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7 |
| Moderator: | NETCAD::COLELLA DT |
| Created: | Wed Nov 13 1991 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 4455 |
| Total number of notes: | 16761 |
While I try to work this further myself, I would like to have any inputs
in regards to this configuration. Will it work?
The idea is to create a duplicate backbone between the VAX and hubs
on different floors, using the spanning tree feature in Bridge1 and 2
to automatically "failover" to the other Ethernet on the VAX in case
of failure of bridge 1 or 2, Ethernet 1 or 2, or DECMR 1 or 2. I'd
like to avoid having to have 2 bridges at each HUB/floor.
My doubts are in regards to the multiple paths existing between the
lower and upper DECMRs, before the bridges have a chance to establish
which one will shut down.
------------- --------------
| B | | | B | |
--| 9 | DETMRs| --| 9 | DETMRs |
| | 0 | | | | 0 | |
| ------------- | --------------
| |
----DECMR-AA ------DECMR-AA
| | |
| ---------------------------+--
| THINWIRE | |
| --------------------------- |
| | |
| | |
DECMR 1 DECMR 2
| |
BRIDGE 1 BRIDGE 2
| |
| |
--------------VAX-------------
ETHERNET 1 ETHERNET 2
Many thanks!
Rolf, EIS Brazil
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 294.1 | Simpler approach? | CGOS01::DMARLOWE | dsk dsk dsk (tsk tsk tsk) | Fri Jul 09 1993 12:37 | 53 |
Are you talking about a dual rail VAX?
If not how about this...
-------------
+ | B | |
[ ]--| 9 | DETMRs|--
+ | 0 | | |
+ ------------- |
+ |Thinwire coax
+ |between HUBs.
+ |
+ |NO management
+ ------------- |cable.
+ | B | | |
[ ]--| 9 | DETMRs|--
+ | 0 | |
+ -------------
+
+
[ ]----VAX
+
If either bridge failed or transceiver cable was disconnected then
the other DB90 would take over. With this setup, the only weak
point is the thinwire coax between the hubs. If that breaks then
both hubs die. Solution for that would be to terminate the hubs and
interconnect them via a 90C repeater in each hub. So if even the
thinwire coax between the hubs broke, the 90C repeaters would segment
and each DB90 would take over in the respective hub. Or since you
have DETMRs, connect 2 90T repeaters together.
-------------
+ | B |9| |
[ ]--| 9 |0| |
+ | 0 |T| |
+ -------------
+ |
+ |
+ |
+ |
+ -------------
+ | B |9| |
[ ]--| 9 |0| |
+ | 0 |T| |
+ -------------
+
+
[ ]----VAX
+
dave
| |||||
| 294.2 | maybe I better scrap it! | VAXRIO::ROLF | Vaporware Design Specialist | Fri Jul 09 1993 16:20 | 23 |
Thanks -.1
Yes, I was thinking in terms of a dual rail VAX for redundancy.
Your ideas are very interesting if I had just TWO hubs, however (and
of course I was not clear about that in my basenote) there will be
some 15 such hubs, not just the two shown, and I believe your approach
wouldn't work then.
Even if I were to use your basic approach, the traffic to all hubs
would go through just one DB90, making ALL the hubs together look
like one BIG workgroup, with more than 200 nodes. Also, I would
exceed the repeater limit, in trying to connect 15 DETMRs together.
With my original idea, I am not sure how a DETMR would react if it
were to see the same packet simultaneously on all of its ports!
Actually it looks like I would be creating a huge loop with every
port sending the same message to every other port at the same time.
Guess I better scrap the idea....
but if anyone has another idea, please share with me!
Rolf
| |||||