| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1893.1 | | NETCAD::STEFANI | Machines to humanize | Thu Dec 07 1995 00:30 | 30 |
| >>1) In dual homing, if we connect M-A and M-B, then M-B link is active
>>and M-A is 'standby', fail over time is in ms, is it correct?
Correct.
>>2) When connecting A-S and B-S, then both link are active and if
>>either A-S or B-S is broken, then the other one will take over all the
>>traffic, is it correct?
If you connect
SAS 1 DAS SAS 2
S----------A B---------S
then you have complete connectivity between nodes SAS 1, SAS 2, and
DAS. If the S-A connection breaks, DAS and SAS 2 keep communicating.
If the S-B connection breaks, DAS and SAS 1 keep communicating.
>>3) A-B and B-A, this is simple, a FDDI dual ring!!!
You got it.
>>4) If I have a DAS devive (A/B ports), if I want to use it as SAS, I
>>have to connect its B port, it is correct? or it depends on the
>>hardware that I am using?
If you have a DAS adapter and you want to connect a single port (A or
B) to a concentrator M port to "act" like a SAS, then it's your call.
You can pick either the A or B port.
- Larry
|
| 1893.2 | some more clarfication, please? | SNOFS1::63496::CHIUANDREW | | Thu Dec 07 1995 20:10 | 15 |
| re .1,
Thanks for your quick answers!
I still have a question on 2) i.e. A-S and B-S:
1) SAS1 and talk to DAS (via A-S) and SAS2 can talk to DAS (via B-S),
then SAS1 and SAS2 can talk via DAS, from SAS1, it just like a direct
connection to SAS2 (what I mean is if SAS1/SAS2 are Gigaswitches (fddi
bridge), DAS is a DECswitch900EF, then DECswitch900EF will not take
part in the FDDI spanning tree calcaulation, is it correct?
thanks again for help!
Andrew
|
| 1893.3 | | NETCAD::STEFANI | Machines to humanize | Thu Dec 07 1995 22:35 | 8 |
| >>connection to SAS2 (what I mean is if SAS1/SAS2 are Gigaswitches (fddi
>>bridge), DAS is a DECswitch900EF, then DECswitch900EF will not take
>>part in the FDDI spanning tree calcaulation, is it correct?
You've hit the limit of my knowledge in this. I'll defer to one of the
GIGAswitch/DEChub folks.
/l
|
| 1893.4 | | NPSS::MDLYONS | Michael D. Lyons DTN 226-6943 | Fri Dec 08 1995 10:36 | 10 |
| Don't think of a-s and b-s as being something special. You have
three interfaces on a ring - that's all that's happening. There is
nothing special with a-s or b-s connections in regards to spanning
tree. The "a" phy/pmd and the "b" phy/pmd are one interface, not two!
The only thing to remember is that if you do the a-s/b-s topology,
you have a wrapped ring. If you lose the system with the das
interface, the others will no longer be able to communicate.
MDL
|
| 1893.5 | | NPSS::MDLYONS | Michael D. Lyons DTN 226-6943 | Fri Dec 08 1995 10:47 | 11 |
| ...in case I wasn't clear - if you have a GIGAswitch/FDDI SAS port (S)
connected to a DECswitch 900 EF DAS port (A) where the DECswitch 900 EF
DAS port (B) is connected to another GIGAswitch/FDDI SAS port (S), the
result is three bridges on one FDDI ring.
What happens in terms of forwarding/blocking behaviour depends on
the rest of the topology. The DECswitch 900 EF DAS port is a single
bridge port - it is not divided into separate bridge ports for the "a"
phy/pmd and the "b" phy/pmd.
MDL
|
| 1893.6 | RE: DECswitch & spanning tree.... | NETCAD::BATTERSBY | | Fri Dec 08 1995 13:06 | 3 |
| And the DECswitch 900EF should take part in spanning tree regardless.
Bob
|
| 1893.7 | what's the rest of the picture? | 35356::RABAHY | dtn 471-5160, outside 1-810-347-5160 | Fri Dec 08 1995 16:41 | 6 |
| re .2:
>... DECswitch900EF will not take part in the FDDI spanning tree calcaulation
Powered, properly functioning bridges always take part in the 802.3 spanning
tree algorithm. It effectively does nothing unless a loop is detected.
|