Satoru-san,
We would like to thank you and the JPOPF for these questions and comments.
If this proposal will be consensus and minimum allocation size to be a /36,
won't the argument that /40 allocations are necessary resurface, ultimately
repeating the same cycle?
One could argue that yes, it can. Having said this, we don't believe that this
will be an issue. A /36 consists of 4096 x /48 subnets, whereas a /32 consists
of 65,536 x /48. Considering that current policy only allows for a /23 IPv4
allocation can be made, even if you were to allocate a /48 to each IPv4 address
in a /23, this would still leave just over 65,000 x /48 subnets unused. This
can be significantly reduced with allocating /36 subnets instead.
The proponent states that unnecessary IPv6 address allocations can be stopped.
However, I understand that the current minimum allocation size was determined
because IPv6 addresses are inherently vast in number, and allocating /32 blocks
was deemed acceptable.
/32 prefixes were deemed acceptable back in 2004, when there were no
restrictions on the size of IPv4 allocations and IPv4 addresses were plentiful.
The inherent vastness of the number of v6 addresses in a /32 are in our view
too great considering the maximum allocatable size.
I recall that reducing administrative overhead was a goal of IPv6 address space
management. While a /32 allocation is expected to require almost no additional
allocations, how are the impacts considered if a /36 is used?
Per the Secretariat's impact assessment, it is believed that there may be an
increased workload while processing requests to return/exchange allocated
space. As for normal applications, we are of the view that reducing the size
from /32 to /36 will create only minimal (if any) additional overheads. Sparse
allocations are already used and would be utilised to ensure effective resource
management of /36 allocations.
It is concerned that this proposal will lead to an increase in IPv6 advertising
routes.
To quote the Secretariat's impact assessment: "We don’t believe that this will
create further fragmentation in the global routing table. The currently
announced assignments would be replaced with the allocated address space". We
agree with this assessment. If a member returns a /32 and announces a /36, this
is a one-to-one replacement and in turn does not increase the number of routes
advertised.
It seems this proposal contains two separate elements: changing the minimum
allocation size and improving the accuracy of WHOIS information. I feel these
should be discussed separately.
We disagree with this assessment. The intent of the proposal is to reduce the
minimum allocation size to assist with maintaining accurate records and prevent
space that will go unused from being allocated.
Regards,Christopher Hawker and Luke Thompson
_______________________________________________
SIG-policy - https://mailman.apnic.net/[email protected]/
To unsubscribe send an email to [email protected]
