[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference netcad::hub_mgnt

Title:DEChub/HUBwatch/PROBEwatch CONFERENCE
Notice:Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7
Moderator:NETCAD::COLELLADT
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

2082.0. "bridge 90, flood mode and more bridges in WG side ??" by STROP::LAYLAND (Iligitami nil Carborundum) Wed Mar 08 1995 15:35

The question concerns wether or not the bridge 90 is considered as a
backbone bridge when in flood mode, ie can there now be more bridges
in the workgroup side of the bridge.

My understanding and that of our UK support centre is that you can not
have another bridge on the workgroup side (including a Brouter 90),
because of the 'spanning tree impementation' in the bridge, and even in
'flood mode' this does not change the rules.

Regards John Layland UK NBU 830-3552
T.RTitleUserPersonal
Name
DateLines
2082.1partly answered off lineSTROP::LAYLANDIligitami nil CarborundumWed Mar 15 1995 14:3816
The question was answered off-line, but there still remains the question concerning
multiple bridges in the same 90 back-plane. Also what of other rules in the docs.




>John, my understanding is that the DECbridge 90 was forced to implement
>full spanning tree because some of our older bridges could not handle
>"leaf bridges" in the tree. If you exceed the 200 node limit on the
>workgroup port, whether by adding another bridge or not, the DECbridge
>90 will go into Flood mode. Note that Flood mode is board rev dependent;
>original DECbridge 90's did not do Flood mode. I think all DECbridge
>90FL's did flood mode.


Regards John
2082.2NPSS::JOHNSONMike J., Network Products SupportWed Mar 15 1995 21:4221
John,

Since I was party to the full text of the off line messages between you and 
Jack, I'll throw in my two cents worth here as well.

There are a couple reasons for not putting multiple DB90's in the same hub.
The first is that there is a very real possibility that someone will form a 
loop, external to the hub, and a subsequent failure, also external to the hub,
will cause ALL of your backbone traffic to transferse the DB90 pair.  
Considering the DB90's performance capabilities, that would be bad news.

Second, when a DB90 "floods" it does NOT act like a "backbone bridge".  It 
actually acts more like a store and forward REPEATER with some degree of
protocol filtering capability.

If you remember and consider the ramifications of just this one thing, the 
limitations become much easier to understand.  The DB90 maintains only ONE
forwarding table, and that table ONLY references the WORKGROUP (hub backplane)
side.

/mj
2082.3STROP::LAYLANDIligitami nil CarborundumMon Mar 20 1995 10:230
2082.4The Brouter is being installed !STROP::LAYLANDIligitami nil CarborundumMon Mar 20 1995 10:5118
Thanks for the reply Mike..  
I understand your comments. There are however a couple of points that need 
clarification.

Some people are putting Brouter 90's in the same backplane as the bridge 90, and 
claiming that it is a legal configuration since v3.9, (flood mode), was introduced.
The other point is if the above is legal, the support centre would claim it as 
illegal since the documentation does not allow it.  

From what you are saying the problems are more likely to be bad configurations than 
a problem of interaction with other bridges.  If that is so then we have the 
classic situation of knowing it works but is unsupported.  The answer to my 
specific situation would be to replace the 90 bridge with a repeater the Brouters 
could now be put into the backplane.  But there will be problem of other sites 
where Brouters and bridge90's are mixed in the same backplane, when/if a fault 
occurs and it is deemed unsupported.

John