RE: new feature in smilint?

["Wijnen, Bert (Bert)" <bwijnen@lucent.com>]

Inline

> -----Original Message-----
> From: Juergen Schoenwaelder [mailto:j.schoenwaelder@iu-bremen.de]
> Sent: dinsdag 2 september 2003 8:44
> To: Wijnen, Bert (Bert)
> Cc: Juergen Schoenwaelder (E-mail); Frank Strauss (E-mail); Mreview
> (E-mail)
> Subject: Re: new feature in smilint?
> 
> 
> On Mon, Sep 01, 2003 at 08:07:24PM +0200, Wijnen, Bert (Bert) wrote:
> 
> > Would it be easy to include a check as to how big
> > a minimum PDU would be if a read-create table needs
> > entries to be created with createAndGo?
> 
> A kind of little programming exercise I would say. Is the assumption
> that all columns which do not have a DEFVAL have to be set? The point
> is that the SMIv2 does not really say which set of columns have to
> be written to create a row and so one has to do wild guesses to get
> an answer to the question you have raised. And of course, we can only
> calculate the PDU size and not the message size since this depends on
> many other factors.
>  
I understand.... so it is only to get a little aid when evaluting.
If you calculate a PDU size of max 200, then I can assume it fits in 484.
If you calculate a PDU size of max 450, then I have to worry and check in
detail.

I would say that maybe you can do
  - PDU size without DEFVAL columns
  - PDU size with DEFVAL columns included

> > More and more people start relaxing the need for 
> > createAndWait support. So we need to check (I think)
> > if a complete row fits in a 484 Octet SNMP Message?
> 
> Perhaps we should just declare 484 dead. :-)
> 
I thought we tried that a year or so ago when we were defining
snmpEngineMaxMessageSize ???  If I remember correctly we were
trying to go for 1472 or something like that as a minimum.
I remember a lot of resistance... 

Bert
> /js
> 
> -- 
> Juergen Schoenwaelder		    International University Bremen
> <http://www.eecs.iu-bremen.de/>	    P.O. Box 750 561, 
> 28725 Bremen, Germany
>