| >How many internal buses does the DECbridge have?
>
>What are their speeds?
>
>Is the bus proprietary?
The bus is proprietary, so answers to these would be misleading
without getting into the architecture.
>Can the diagnostics be performed while the bridge is
>on-line?
No.
>How long does it take the bridge to reboot ?
Approximately one minute.
>Can the bridge be prevented from accepting a downline
>upgrade, ie by a DIP switch ?
Yes -- both by means of a switch on the hardware switchpack and
a management-settable switch (via SNMP or ELMS).
>Does the bridge do LAT traffic compression ?
No, this is appropriate only for point-to-point, remote links, not for
local bridges connecting LANs. The DECbridge models don't have
point-to-point links (Hastings does however).
>Is either the bridge or any of its modules fault
>tolerant?
Our bridges are inherently fault tolerant due to the spanning tree
algorithm. (Backup bridges take over in case of failure of other
bridges.)
>Are any of the modules hot swapable?
No.
>Are the following Spanning Tree parameters settable?
> Bridge Priority
> Port Priority
> Hello Time
> Maximum Age
> Forward Delay
> Long Cache Timer
> Path Cost
> Port Cost
Yes except the port priority. There is no benefit to changing the
port priority. Also, the Path Cost is determined by Spanning tree
based on the dynamic tree configuration (it's defined as the
cost of this port to the Root); it's only the Port Cost that
can be manipulated in any bridge.
Anil
|
| re: .1
>>Is either the bridge or any of its modules fault
>>tolerant?
>
> Our bridges are inherently fault tolerant due to the spanning tree
> algorithm. (Backup bridges take over in case of failure of other
> bridges.)
Gee, Anil, with an answer like that, you have a great future in
marketing...:-)
None of the DECbridges are fault-tolerant in the 3000ft sense. If one
of the four logic modules goes south, the bridge will fail self-test,
and not go online.
However, all DECbridges can tolerate "external" datalink failures
(transceiver failure, unterminated segment, backhoe fade, etc.) and
will continue to operate on the unaffected port(s). This is of limited
value in a two-port bridge, but certainly a worthwhile feature in the
three- and four-port variants.
|
| No marketing for me, thanks..
One of the two main goals of the spanning tree algorithm is
fault tolerance. In simple devices such as bridges, the meaning
of "fault tolerance" can be nebulous; in this case there seemed to
be a clear answer. Yes, everybody does spanning tree these days
but why not get across a positive message when we can?
Anil
|