[address-policy-wg] Policy proposal: #beta: IPv4-HD-Ratio

On 26-apr-2005, at 15:22, Michael.Dillon at radianz.com wrote:

>>> Note that the current HD ratio for all IPv4 address space that isn't
>>> reserved by IANA is 90.45%.

>> As you see the HD Ratio propsed is much higher but
>> would help LIR's with bigger allocations to justify
>> their IP usage.

> You are comparing apples and oranges. Or maybe you have painted
> all your apples with orange paint.
>
> The proposal suggests that the HD ratio should be 0.96
> It does not mention a percentage. In fact, a percentage
> is a kind of ratio but it is not the same kind of ratio
> as the HD ratio.

 From RFC 3194:

               log(number of allocated objects)
    HD = ------------------------------------------
          log(maximum number of allocatable objects)

    This ratio is defined for any number of allocatable objects greater
    than 1 and any number of allocated objects greater or equal than 1
    and less than or equal the maximum number of allocatable objects.
    The ratio is usually presented as a percentage, e.g. 70%.  It varies
    between 0 (0%), when there is just one allocation, and 1 (100%),  
when
    there is one object allocated to each available address.

> The comment regarding all IP address space that is not
> reserved by IANA is not clear whether it is talking about
> an HD ratio or an allocation percentage. And the most important
> thing is that it does not say what is the source of the
> numbers that lead to the 90.45 result.

Out of the 256 /8 blocks 32 are class D (224 - 239, multicast) and E  
(240 - 255, reserved) and three others are also unusable: 0, 10 and  
127. Of the 221 usable /8s 72 were unused as of March 2005. See  
http://www.iana.org/assignments/ipv4-address-space . According to the  
ISC Domain Survey at http://www.isc.org/ds/ 317 million IPv4 hosts  
had a domain name in January 2005.

So that's 149 * 2^24 = 2.5 billion allocatable objects with 317  
million objects allocated, or: 8.501 / 9.397 = 0.9047. (log base 10.)

> HD ratio for IPv4 is intended to count the number of addresses
> assigned by an LIR and compare that to the number of addresses
> that RIPE (or another RIR) has allocated to the LIR.

No, the intent of the HD ratio is to show that we need IPv6 even  
though only some 9% of all usable IPv6 addresses are in use.

(Throw in standard disclaimer about the host count.)

> When discussing
> "all IPv4 address space" or "IANA reserved address space" we are
> talking about address attributes that are not covered by the
> HD ratio as we know it.

The HD ratio is just a rule of thumb that says on everage, we waste  
one fifth to an eighth of the address length. The fact that we're now  
apparently using more than 90% oof the address length while including  
sparsely populated pre-1993 address space shows that the HD ratio  
isn't all that useful, although it's main point, that there are  
losses at allocation boundaries, is of course very true.