Re: [Isms] IsmsQuestions on TLSTM/TSM and RFC 5343

[Wes Hardaker <[e-mail]>]

>>>>> On Sun, 23 Jan 2011 18:24:33 +0100, Frank Fock <[e-mail]> said:

FF> That's why I thought about simplifying
FF> the SNMPv3 and making the discovery of
FF> a contextEngineID unnecessary if not
FF> required by the command generator.

>From a pure architecture point of view, I agree.  Though I don't think
it would be wise of a SNMP stack to require the "actual" end-application
to think about contextEngineIDs if it didn't have to.  I think most
stacks would better server their users if they had an API that handled
it all, regardless of where the contextEngineID discovery happened.

EG, hypothetical pseudo fake code:

snmpv3_response *
send_snmpv3_request(struct snmpv3_session *session, struct snmpv3_pdu *pdu) {
  if (session->contextEngineID == NULL) {
     snmpv3_response *response =
         send_pdu_to_dispatcher(session, new_contextEngineID_probe_pdu());
     session->contextEngineID = extract_contextEngineID(blob);
  }
  return send_pdu_to_dispatcher(session, pdu);
}

That seems a more helpful API than making each application doing the
steps internally to that.
  
-- 
Wes Hardaker
Cobham Analytic Solutions