Re: b/w example values

[Wes Hardaker <wjhns1@hardakers.net>]

>>>>> On Wed, 3 Mar 2004 15:47:04 -0500, Bobby Krupczak <rdk@krupczak.org> said:

>> Now, that I agree with.  We're not going to come to agreement as our
>> needs differ.

Bobby> The discussion over defaults/not-defaults is an engineering and
Bobby> design tradeoff with no realy discernible difference to our
Bobby> end-users (I hope).

Our needs differ.  My end-users really do care about bandwidth and cpu
issues.  In fact, they care so much that it's unlikely they'll even go
with this protocol in the first place unless the compression really
does make it close enough to binary encoded protocols that they'll be
happy.

Our needs differ, because our customers differ.

Bobby> I seriously asking you if you really think that defaults/no-defaults
Bobby> will effect adoption?

Yes.

Bobby> I just cant see that it matters to end-users.

Not to yours.

Bobby> What environments are you working in that you think the needs are
Bobby> drastically different?  I think this discussion is important.

low-power, low-bandwidth, low-connectivity,
frequently-dropped-packets, ...

[in various combinations.  Sometimes in all]

[(battery) power actually is the hardest to accommodate, since it's
affected by both CPU and bandwidth for the same data so it takes a
double hit much of the time]

Bobby> In the case of your cidrTable example, 5 -1s (not specificed because
Bobby> they are defaults) only amounts to :

Bobby> 5 * (1-byte tag, 1-byte len, 4-byte value) = 30 bytes

30 bytes * 10000 routes * 10000 hosts to distribute it to
   (ok, that's too extreme... but you get the point (maybe)...)

However, in xml that encoding is quite a bit bigger by the way if you
include the tags:

% foreach i (1 2 3 4 5)
  echo "<ipCidrRouteMetric$i>-1</ipCidrRouteMetric$i>" >> data
  end
% wc -c data
    220 data
% gzip data
% wc -c data.gz
     81 data.gz

(The ascii encoding is > 2.5 times larger even when compressed.)

Bobby> I hardly consider an extra 30-bytes xmit'ing over a 10Mbps or the more
Bobby> likely 100Mbps anything to sneeze at.

Would you believe some of my environments deal with 128kb.  Heck, some
are still using 50 baud modems (but fortunately, I don't need to worry
about them).

Bobby> Here is the blurb from the KB regarding SNMP polling:

[btw, I don't think we need to argue about how lame SNMP is at
encoding data too...  It has its drawbacks as well that are very well
known.  Many papers have proven this, but we're not discussing SNMP here] 

Bobby> Does this help?

Not me.  It still lets me know our target operating environments
differ (I'm not sure why you're trying to convince me otherwise.  I
have no problem admitting that in your large bandwidth environment you
have no problems with adding as much data as you can.  I agree, it
probably isn't a problem for you.  You're welcome to send all data
including defaults with your data because you don't care.  Mandating
it into the standard means you affect me, however, and the
environments I operate in which are different).

-- 
"In the bathtub of history the truth is ha

--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>