| Your understanding is correct; you can do either.
Most devices are managable via the hub IP address (notable exceptions:
DECbridge90's, all DECservers, DECbrouter90s, WANrouter90, MUXserver).
Using the IP address of the hub is the only way to get Out of Band
Management of the hub devices.
Obviously, you save lots of addresses.
There is some performance penalty for using the HUBs address, but I
don't have any numbers for that.
Note that current versions of HUBwatch don't make any attempt to
optimize the choice of IP addresses; if a device is in a hub and you
select it by double clicking on it in the hub, then you'll use the hubs
address even if it has one of its own - you need to switch agents
explicitly using the agents table to use its own address.
|
|
Add all currently shipping Token Ring modules and the DECpacketprobes
to the list of exceptions. A new TR mau is coming, and should allow
either its own or the hub's IP address to be used. The DECpacketprobes
require you to use the module IP address. The DECagent 90 is also
on the exception list, although I'm not sure why anyone would want
to try to manage it using another IP address.
This has to do with backplane management protocols supported by a
module; some modules just support the minimum (older 90 modules
and most 3rd party modules are like this).
|