number of unaggregated class C's in swamp?

Alan Hannan alan at
Mon Oct 2 05:19:41 UTC 1995

.........  Yakov Rekhter is rumored to have said:
] So, perhaps we should just look at the total amount of IP address space
] advertised by a provider in its routing advertisements, then divide 
] this amount by the number of routes the provider advertises, and 
] see whether the resulting number meets the goal.

  But what is the goal?  

  I thought about your jottings all weekend Yakov.  I'm sure it will comfort
you to know I thought of you over the weekend :)

  My opinion is that the Internet became a classic chaos model directly
after the transition from a single NFSNET backbone to the NAP model we now
live with.  At this point in time we are debating and proposing the methods by 
which we will impose laws on this system.  As one studies chaotic models, one 
often finds that natural laws tend to impose possible goals called 'attractor 
points'.  I believe our goal for 'attractor points' would be that people have 
a respectable ratio of networks announced to host space claimed.  However,
I haven't the foggiest how to define this goal.  Do we factor in the amount
of address space a person has?  Should this be a factor?  But how do we factor
in percentage usage, and future growth?  These are the questions that have
been kicked around always.

  My capitalist nature says that the amount of address space one has should
not be an issue.  I'm not terribly sure on how that enters into the metric.
I'd be in favor of something that directly associated 'goodness' or 'cost'
with the amount of ip nodes one could route, or the ratio of routes to nodes.

  I attempted to do some work on this, and basically came to no conclusion. 
Perhaps my futility will spawn some other, more intelligent thought.  First,
I took the radb from and
parsed it into a database that looks like this:

AS Origin:Number of Routes:Address Space (ie #of ip nodes):nodes/routes

  I then started using 'gnuplot' to compare fields to fields, looking for some
order in the chaos.  Surprisingly, I found none.  If you're interested in the
databse, it's available from
It may be that I hosed my math up in some of the places, because some of my
numbers look kind of wacky.

  So, here are a few of the graphs, in hope that someone might comment on
them, or gain some sort of insight.....

  Here I plotted the number of routes vs the metric Yakov mentioned, hoping
to see some sort of a 'sampling' of where we are.  However, we don't really...
Or do we?

   100000 ++-----------+-------------+------------+-------------+-----------++
          +            +        A    +            +             +            +
    80000 A+                                      routes vs nodes/route  A  ++
          AA  A                                                              |
    60000 A+       A                                                        ++
    40000 A+            A                                                   ++
          AAA                                                                |
    20000 AAA A A                                                           ++
          AAAAA AA A   A             +   A  A     +        A    +A           +
        0 AAAAAA-A-A-A-+--A----------+----------A-+-------------+-----------++
          0           500          1000         1500          2000         2500

  Well, we seem to see some sort of correlation where the more routes a person
has, the lower their metric becomes where metric = hostspace/routes.

  Interestingly, I found that the older an AS, the more routes they had.  Big
surprise there.... :-).

Number of Routes
     1000 ++--------+---------+---------+---------+---------+---------+-----++
          +         +         +         +         +         +         +      |
      800 ++    A                                    AS# vs # of Routes  A  ++
          |                                                                  |
      600 +A           A                                                    ++
      400 ++      A           A                                             ++
          AA                   A    A                                        |
      200 AAAA         A     A         A    AA                              ++
          AAAA  A A +AAA   AAAAA  AAA AA+ AAAAAA A+ A       +         +      |
          0       1000      2000      3000      4000      5000      6000
                                         AS #

  And finally, I found that the older an AS number, in general, the more IP 
space they had, at least to some degree:

Address Space
    1e+07 ++--------+--------+---------+--------+---------+--------+--------++
          +         +        +         +        +         +        +         +
    8e+06 ++          A                            AS# vs Address Space  A  ++
          |                                                                  |
    6e+06 ++                                                                ++
    4e+06 A+      A          A        A                                     ++
          |A                      A       A                                  |
    2e+06 +A   A      A    AA A   A            A                            ++
          +AAA A A  + A     AAA A AA AA+  A A A +         +        +         +
          0       1000     2000      3000     4000      5000     6000      7000
                                         AS #

  Well, perhaps these were interesting, perhaps not.  Perhaps they'll help
people think of a brilliant wiz bang solution to this whole huge routing table
mess.  Most likely not.  Oh well.

Alan Hannan                                          Email: alan at
Network Systems/Security                             Voice: (402) 472-0239
MIDnet, Lincoln NOC Office                           Fax:   (402) 472-0240

 " The reasonable man adapts himself to the world; the unreasonable one 
   persists in trying to adapt the world to himself. Therefore all progress 
   depends on the unreasonable man. "  -  George Bernard Shaw 

More information about the NANOG mailing list