bducharme AT avdl DOT com writes:
> Can someone answer me,
> 
> How Tivoli Netview do his dicovery in his network ?

Someone more qualified than me is bound to respond, so I'll put up a
strawman.  

First, discovery can be limited/directed by the inclusion of a
seedfile.  The server-setup command or menu item, or netmon config
screen in "smitty nv6000" (assuming AIX) specifies the file location.

/usr/OV/conf/netmon.seed is a typical name used in our shop.

>From there, arp caches are the key, starting with your own machine's
arp cache and then those in the seedfile (depending on autodiscovery
settings).

> Pooling or receiving everything ?

Not sure what that means. 

> By snmp ?

SNMP is the key.  With autodisvoery turned on, they query the arp
cache portion of MIB-II in those devices and branch out from there. A
common tactic is to put your main routers in the seedile, and discover
out fro mthere. 

> By icmp ?

ICMP is used as a precursor to SNMP access attempts, i believe. 

> How can we block him from discovering networks from Internet that don't
> belong to us ?

Limiting discovery by appropriate use of the netmon seedfile.  Check
your netview administrator's guide for tips tricks and how-to's on
seedfile setup. 

You can also turn off autodiscovery and tell netview only to discover
the things you call out in your seedfile.  You can specify address
ranges to exclude, include, and all sorts of goodies. 

--