[LINK] IPv4 host density measured by ping

[Robin Whittle]

Here are some other /24 prefixes nominated by people who wrote to me
directly:

  150.101.117.0/24  Internode                    35 acks

  Only has a few reverse-mapped names:

   smtp1.abigroup.com.au.
   gi0-213.cor1.syd6.internode.on.net.
   bdr1.netro.com.au.
   bdr1.netro.com.au.
   bdr1.syd.bigdy.net.au.
   gi0.105.cor1.syd7.internode.on.net.


  59.167.26.0/24   Internode - "dynamic" I am told.  93 acks

   All reverse mapped in the form of:

     ppp26-1.lns1.syd6.internode.on.net


I think Netcraft only surveys web servers:

   http://news.netcraft.com/archives/web_server_survey.html


The last ping survey I know of was in 1999.

  http://www.isc.org/index.pl?/ops/ds/reports/2006-07/

This was not a random scan of the Net, as I did - though I also scanned
particular /8 prefixes, and sets of BGP prefixes of different lengths.
They scanned the DNS for hostnames with IP addresses (mainly web server
names, I guess) and then pinged those IP addresses.

           |     Survey    Adjusted     Replied
     Date  | Host Count  Host Count    To Ping*
   --------+-----------------------------------
   Jul 2006|439,286,364                  -
   Jan 2006|394,991,609                  -

   Jan 1999| 43,230,000               8,426,000

   [* estimated by pinging a sample of all hosts]

I don't know of any ping surveys, so wrote software to do my own.

I know pings are a limited indication of IP address space usage.  I
guess they generally give a lower bound.

I recognise this doesn't count the number of computers behind a NAT
firewall - I am not attempting to count those.

I wanted a rough estimate of how many IP addresses are genuinely being
used at present.  I also wanted to find trends in IP address
utilisation.  For instance, I found the 19 big /8 prefixes advertised in
BGP being very poorly utilised, if 0.2% ping responses is anything to go
by, when others are more like 10 or 20%.

 - Robin       http://www.firstpr.com.au/ip/host-density-per-prefix/