Re: A proposal about hostcount and DNS

[Magnus Danielson <]

>>>>> "MN" == Mike Norris mnorris@localhost writes:

Hi Mike!

 MN> Magnus
 MN>       thanks for your thoughts and suggestions.  They
 MN> are worthy of discussion in other lists too, but you're
 MN> probably right to start with the experts in the DNS
 MN> WG, of which I am not one.

Mirjam did ask me to drop a proposal at DNS WG. Obviously she thougth
there where some content in it...

When I wrote the mail I did leave some things out and some of that has
obviously created questions.

I did assume that people would update their hostcount number in time
for the hostcount (at the end of each month) and that this hostcount
represents some number close to (but not necessarilly at) the reality.

I did assume that a definition of the word hostcount (or which ever
term would be prefered) would be done in some document so that no
alternative interpretations would arise on a wide scope.

 MN> I don't think that at any time the DNS host count has
 MN> ever pretended to give even an approximate measure of
 MN> the number of Internet hosts, let alone the number of
 MN> Internet users.  For example, there are some hostmasters
 MN> who use DNS as a database for all systems in their
 MN> jurisdiction, whether such systems have Internet access
 MN> or not; this has been going on for a long time.  More
 MN> recently, there are many Internetters using machines hidden
 MN> behind application firewalls and filter lists which do not
 MN> appear in the host count.

Well, I have personally never assumed much correctness from the
hostcount myself, rather a good hint on where things is moving.
However, as new people enter the stage they may not even take
the time to understand the missing pieces and therefore could some
closer numbers be of interest. This is not a main point, but could
have some influence.

 MN> Perhaps the main use for the host count has been its
 MN> recording of growth in individual countries (and gTLDs),
 MN> in regions and in the world.  It's really the change and
 MN> the rate of change that we get from the host count; the
 MN> absolute numbers are not that meaningful.

Absolute/Exact numbers is not the goal, just numbers which is closer
in some sense. Maybe will this hidden factor have a greater influence
than we think and therefore would the hostcount be somewhat false
numbers.

 MN> There are many factors causing the increase in the host
 MN> count.  These include more desk-top computers (and less
 MN> multi-user systems), more ISPs, more connectivity, etc.
 MN> These factors and others seem so far to have outweighed
 MN> the advent of firewalls, as the curve continues inexorably
 MN> upwards.  Of course it would be nice to know the number
 MN> of 'hidden' hosts, and this could only be done on a 
 MN> trusted basis and with full cooperation from people at
 MN> all levels - not an easy task.

No, it is not a easy task, but maybe is there enougth interest in
doing this.

 MN> If this could be done, do you think there should be two
 MN> separate hostcounts - visible and hidden?  If we combine
 MN> both in a single figure, we lose important information and
 MN> also change the meaning of historic data - there will be
 MN> a discontinuity in the curve.

I agree that one cannot just change method overnigth and I also agree
that just cludge the visible and hidden number together would be to
loose interesting information. If it is feasable to do a separate
report of hidden and visible hosts then I think it should be done.

Cheers,
Magnus