Newsgroups: comp.protocols.snmp
Path: nighthawk.clearpoint.com!nero!martillo
From: martillo@nero.clearpoint.com (Joachim Carlo Santos Martillo Ajami)
Subject: Re: What NOC's support SMP?
Message-ID: <1993Feb2.083913.26103@nighthawk.clearpoint.com>
Sender: usenet@nighthawk.clearpoint.com (USENET News System)
Nntp-Posting-Host: nero.clearpoint.com
Organization: Clearpoint Research Corp., Hopkinton Mass.
References: <1993Jan30.165240.15098@nighthawk.clearpoint.com> <C1rzqC.6yF@apertus.com>
Date: Tue, 2 Feb 1993 08:39:13 GMT

In article <C1rzqC.6yF@apertus.com> stuarts@news.apertus.com (Stuart Stanley) writes:
>Joachim Carlo Santos Martillo Ajami (martillo@nero.clearpoint.com) wrote:
>: In article <karl.9.728359086@empirical.com> karl@empirical.com (Karl Auerbach) writes:

>: Let's all look up the word profiteering in the dictionary.

>Lets also look up the idea of "it works".  

Let's look up the word "non-comprehension."  The existence of a few
public domain versions of SNMPv2 proves nothing.  The problem lies in
all the embedded SNMPv2 systems which now must be developed not to
mention maintaining the existing incompatible base of embedded SNMP
systems.

Oh, Case says "no problem," you can just run proxies.  Well, what if
proxies are not available in the target environment?  Also the proxy
becomes an extra unneeded point of failure.  And last, who is going to
debug the obscure associated problems?  The proxy supplier or the
embedded system manufacturer?

In another context I pointed out the following.

"Writing a program or designing a board is easy.  Creating an
organization which can identify or create the market, get the product
to market, support customers and put together the system wherin the
financial institutions, the corporation and the customers can all
interact to every one's mutual benefit is the really impressive
achievement.  A person who doesn't understand this reality doesn't
understand anythin (BTW -- like most engineers) and certainly doesn't
understand systems or system architecture in any way, shape or form.

The retrofitting of ISODE or CMU-SNMP with SNMPv2 prove nothing
whatsoever.  The claim that such retrofitting proves anything proves
only a complete lack of understanding.

>					    After that we can look up
>the idea of "it works perfectly for the needs we have".  The only
>problem we ever had with SNMP was is inherent lack of security.  This
>has been addressed to a fair degree with SNMP V2 and thus the whole
>protocol fits our needs to the T.  I guess that is why the Internet
>community (or at least this prarticular fragment of it) are having
>SNMP foisted on it.... by itself!

See the appended e-mail.

>In any case, it is a working, accepted technology that in theory will be
>replaced by CMIP some day.  Dead dog with fleas, fine!, but I still need
>a dead NOC with fleas as well ;)

SNMP is accepted because Case, Rose and their ilk hyped it to the
point of saturation.  Now there is an installed embedded base.  Well,
it is too late.  Either SNMP was a product or it was an experiment.
If it was an experiment, it should never have been made into a
requirement.  If it is a product, the IETF should live with it.

Otherwise, I have to wonder about other IETF requirements.  Will they
change out from under the installed product base?  Is the IETF being
used to engage in crude market manipulation?  There is a term for this
practice.  The term is "restraint of trade."  There are a lot of people
who have serious problems with attempts at restraint of trade.

>: Joachim Carlo Santos Martillo Ajami

The following e-mail may be of interest.

From {3COM/PDD/PeteW}@pdd.3mail.3com.com Tue Feb  2 02:48:22 1993
Date: Mon, 1 Feb 93 19:28 PST
From: {3COM/PDD/PeteW}@pdd.3mail.3com.com
Subject: Re: Squashing important ideas
To: snmp2@thumper.bellcore.com, snmp-sec-dev@tis.com
Msg-Date: 1993-02-01
Msg-Time: 19:22

Microsoft Mail v3.0 IPM.Microsoft Mail.Note
From: Wilson, Peter
To:  Secure SNMP
     SNMPv2
Cc:  Heads, Bob
     Woodruff, Paul
Subject:  Re: Squashing important ideas
Date: 1993-02-01 19:21
Priority: 
Message ID: D79ABD08
Conversation ID: D79ABD08

-----------------------------------------------------------------------------

I've been fairly quiet on these mailing groups since Christmas. The
reason?  To follow from Cheryls message a few days ago, I'm also
DISGUSTED by the way the work of this group has been carried out. I
was at the Knoxville meeting when I think it was Jeff Case said
something along the lines of 'We've done our bit with SMP, now we're
turning it over to the community'! This never actually happened. The
original authors effectively kept complete control over the whole
process. Example:

One of the points Cheryl made was that ideas should be dealt with on
an equal basis. This has patently not happened in the SNMPv2 group.
The bias is obvious. If you are one of the original authors silence
means acceptance of a proposal, if you're not then silence means
rejection! The original authors are NOT now the only experts in
network management. They may not even now be the best experts. A whole
body of experience has been totally wasted in achieving these
standards.

THE RESULT
============

So, what is SNMPv2, the thing we've all worked so hard for? Its Secure
SNMP except:

1) All the security is optional so it won't be adopted by many people.
Even if they do the user will never understand it!

2) An even more complex administrative model that, coupled with the
optional security, means everything it does could be achieved with
community strings in SNMPv1

3) Of course, not forgetting.....

	The Awesome 'Get-Bulk' PDU which is less effective than it
could be because the available packet space, after heaers, is smaller
and which a user should never even see!

4) A few trivial changes to MIBs, just to make them incompatible and
take twice as long to produce.

5) Oh yes, we all get one more marketing 'tick-in-the-box' and a few
people can make lots of money writing books on how different SNMPv2 is
over SNMPv1 and exactly why a customer should look for that latest
buzz word.

Chuck is actually right to say what are the requirements.

WHOSE FAULT
===============

I can't actually blame the authors for trying to get everything their
own way. Wouldn't you? The blames list largely on the WG chairs and on
the majority of the 150(?) people in the SNMP meetings and the people
registered with these groups. If your on this group and aven't taken
an active part then you deserve SNMPv2!

So thats why I've been quiet. NOT BECAUSE I AGREE WITH EVERYTHING
GOING ON, NOT BECAUSE I THINK SNMPv2 IS WONDERFUL but because there
are much better brick walls to bash my head against than having to go
all the way to the IETF!

I've given up on SNMPv2, its a farce. I'll sit back, wait a year or so
for the next IETF management protocol and hope the IETF organises
itself better.

Pete










-- 
Joachim Carlo Santos Martillo Ajami

The article represents noone's opinions other than the author's.