[ppml] HD Ratio changes

Charles Scott cscott at gaslightmedia.com
Wed Feb 11 10:12:01 EST 2004


Michael:
  I believe the simplified formula for calculating the number of hosts for
".996 HD" in Perl is.

    exp(log($total_hosts)*.996)

  In any case it gives the same numbers your code does. From this it
appears that if some HD ratio is adopted that the .996 number may not be
appropriate.
  Also, since we're talking about an allocation/assignment policy, 
shouldn't the HD ratio be applied to the total volume of addresses that 
have been so assigned rather than the count of hosts deployed in that 
address space? I wonder how that would affects the results.

Chuck Scott



On Tue, 10 Feb 2004, Owen DeLong wrote:

> OK... I took your mail, and tried to make this simple...
> 
> I wrote the the attached PERL script to compute the number of hosts needed 
> for
> a .996 HD ratio and the number of hosts needed for 80% utilization and came
> up with the following table.  Either HD doesn't do what you claim, or, it
> isn't simple enough for me and my math didn't work right.  The table here
> shows that in all cases, a .996 HD ratio is a larger number than the
> 80% requirement, and, that the more space you have, the worse it gets.  I'm
> guessing I've got an error in my code, but, this is as easy as I thought it
> could be to compute the HD ratio needed for a given block size.  If you have
> an easier way, please post code.
> 
> 
> +--------+-----------+-----------+-----------+---------------+
> | Prefix |  N Hosts  |  .996 HD  |   80% HD  | .996 HD - 80% |
> +--------+-----------+-----------+-----------+---------------+
> |     /3 | 536870912 | 495393603 | 429496729 |      65896873 |
> |     /4 | 268435456 | 248384517 | 214748364 |      33636152 |
> |     /5 | 134217728 | 124537069 | 107374182 |      17162886 |
> |     /6 |  67108864 |  62441419 |  53687091 |       8754327 |
> |     /7 |  33554432 |  31307393 |  26843545 |       4463847 |
> |     /8 |  16777216 |  15697157 |  13421772 |       2275384 |
> |     /9 |   8388608 |   7870369 |   6710886 |       1159482 |
> |    /10 |   4194304 |   3946111 |   3355443 |        590667 |
> |    /11 |   2097152 |   1978533 |   1677721 |        300811 |
> |    /12 |   1048576 |    992013 |    838860 |        153152 |
> |    /13 |    524288 |    497383 |    419430 |         77952 |
> |    /14 |    262144 |    249383 |    209715 |         39667 |
> |    /15 |    131072 |    125037 |    104857 |         20179 |
> |    /16 |     65536 |     62693 |     52428 |         10264 |
> |    /17 |     32768 |     31433 |     26214 |          5218 |
> |    /18 |     16384 |     15761 |     13107 |          2653 |
> |    /19 |      8192 |      7901 |      6553 |          1347 |
> |    /20 |      4096 |      3961 |      3276 |           684 |
> |    /21 |      2048 |      1987 |      1638 |           348 |
> |    /22 |      1024 |       995 |       819 |           175 |
> |    /23 |       512 |       499 |       409 |            89 |
> |    /24 |       256 |       251 |       204 |            46 |
> |    /25 |       128 |       125 |       102 |            22 |
> |    /26 |        64 |        63 |        51 |            11 |
> |    /27 |        32 |        31 |        25 |             5 |
> |    /28 |        16 |        15 |        12 |             2 |
> |    /29 |         8 |         7 |         6 |             0 |
> |    /30 |         4 |         3 |         3 |             0 |
> |    /31 |         2 |         1 |         1 |             0 |
> +--------+-----------+-----------+-----------+---------------+
> 




More information about the ARIN-PPML mailing list