RE: SNMP question

["Harrington, David" <dbh@enterasys.com>]

Hi,

I don't quite agree that the manager is a PDP and the agent a PEP.

If the application makes the decision about which policies should be
distributed, then the application is a PDP. If a device enforces the
policies, then the device is a PEP. 

SNMP can be used to distribute policies from the PDP (manager) to an
SNMP agent (on the PEP), and the MIB is a schema to specify the policy. 

You also need to be careful thinking that all SNMP entities are either a
manager or an agent. IN SNMPv1 that was true; in SNMPv3, that
distinction is much less clear. Each SNMP entity can provide "agent"
types of functionality and can provide "manager" types of fucntionality,
in a variety of combinations.

dbh


> -----Original Message-----
> From: Wijnen, Bert (Bert) [mailto:bwijnen@lucent.com] 
> Sent: Sunday, January 11, 2004 5:57 PM
> To: chum@noos.fr; ops-nm@ops.ietf.org
> Subject: RE: SNMP question
> 
> If you look at the DIFFSERV MIB (RFC3289), and if you
> take into account that that MIB representes the management
> objects for Diffservv at a managed device (or SNMP agent
> if you like) then indeed you interpretation to consider 
> an SNMP manager as PDP and an SNMP agent as PDP (in the
> context of these tpyes of MIB modules).
> 
> Thanks,
> Bert 
> 
> > -----Original Message-----
> > From: c-h u-m [mailto:chum@noos.fr]
> > Sent: zondag 11 januari 2004 5:20
> > To: ops-nm@ops.ietf.org
> > Subject: 
> > 
> > 
> > Hello,
> > 
> > I want to ask some questions in the snmpconf wg but I cannot 
> > subscribe to the mailing list of this group. So, I ask this 
> > quesion here.
> > 
> > I am studying Policy-based management and trying to mapping 
> > SNMP entities to PBM entities. Could somebody confirm my 
> > understanding that SNMP manager corresponds to the PDP 
> > (Policy Decision Point) and SNMP agent corresponds to the PEP 
> > (Policy Enforcement Point)? I hope that is right and it is 
> > simple to see that but I cannot find any document which 
> > explicitly states that.
> > 
> > Thanks,
> > 
> > 
> > 
> 
>