Re: SNMPv1 and Reverse Poll (managed ->manager)

karl@mel-brooks.empirical.com Wed, 13 October 1993 03:01 UTC

Date: Tue, 12 Oct 93 14:20:42 PDT
Message-Id: <9310122120.AA11219@mel-brooks.empirical.com>
To: mike@premisys.com
Cc: snmp@lists.psi.com
In-Reply-To: Mike MacFaden's message of Tue, 12 Oct 1993 13:33:30 -0700 (PDT) <9310122033.AA29462@premisys>
Subject: Re: SNMPv1 and Reverse Poll (managed ->manager)
Reply-To: karl@empirical.com
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: karl@mel-brooks.empirical.com
Repository: empirical.com
Originating-Client: mel-brooks

 >   > "What we need is somthing that has these properties:-
 >   > - An ability to originate management reports from the managed node (i.e.
 >   >         eliminate polling).
 >
 >   In SNMPv1 it is quite possible to include a mib object that defines a set of
 >   Enterprise specific traps that contains all the information you'd ever
 >   want a node to report to an NMS and have it sent to the NMS based on some 
 >   mib object interval.
 >
 >   I don't think this goes against any RFC, please correct me if I am wrong!

It is completely acceptable practice, both in V1 and V2.

It is also is acceptable practice in both V1 and V2 to temporarily
reverse agent/manager roles and let the normally-agent become a
manager for a moment and let the normally-manager become a momentary
agent and send a set-request PDU with its confirming response.  In V2,
however, there is a better way.

			--karl--

SNMPv2's properties MARTIN USHER
Re: SNMPv2's properties Bob Stewart
SNMPv1 and Reverse Poll (managed ->manager) Mike MacFaden
Re: SNMPv1 and Reverse Poll (managed ->manager) karl
Re: SNMPv2's properties Aleksey Y Romanov