Wasted space

While importing netblock data I discovered what must be common
knowledge for most. A large portion of allocated space is not being announced.
Equally exciting was the discovery that a large portion of space is
maintainer-free, thus making it essentially impossible to have an exact
accounting of who has what.
  IMO this seems much like those people who hoard domain names, except
worse because ipv4 space is finite.
  Seems like we ought to drop these pointless allocations and make the
owners justify space they use in the future. Networks using such space for
internal use can use NAT if they ever decide to connect the The Internet.

  Austin

Equally exciting was the discovery that a large portion of space is
maintainer-free

in the irr or (arin|ripe|apnic)?

much of the the former thanks to sean and to baby peter. the latter should
not be the case.

IMO this seems much like those people who hoard domain names, except
worse because ipv4 space is finite.

domain names are finite. 254 char limit drawn from an 8-bit char set.

randy

IMO this seems much like those people who hoard domain names, except
worse because ipv4 space is finite.

domain names are finite. 254 char limit drawn from an 8-bit char set.

$ bc
256^254
49311837877366649323600580884811328064642490645928167773636391338386\
00942820417921935608125537553934278674005267623599165972833122328326\
58311281622107670335702985799671951234310153163915857728680359766210\
69439038508288907840911493166867209378778336289339669574030006474132\
65364309855012299736389026478635486131947843882498538312526670313197\
24958132568898411896638150110768600863536200871492771279798342546336\
76061407041110011837155687183077462622686306172536143846476937385117\
82868915581833149250995402477804959206649465186461985527496130098804\
49926596639031121858756000207590413184793166384097191709192063287296

256^4
4294967296

  Austin

domain names are finite. 254 char limit drawn from an 8-bit char set.

$ bc
256^254
49311837877366649323600580884811328064642490645928167773636391338386\
00942820417921935608125537553934278674005267623599165972833122328326\
58311281622107670335702985799671951234310153163915857728680359766210\
69439038508288907840911493166867209378778336289339669574030006474132\
65364309855012299736389026478635486131947843882498538312526670313197\
24958132568898411896638150110768600863536200871492771279798342546336\
76061407041110011837155687183077462622686306172536143846476937385117\
82868915581833149250995402477804959206649465186461985527496130098804\
49926596639031121858756000207590413184793166384097191709192063287296

256^4
4294967296

your point?

your point?

  It would be possible to assign names to every atom in the universe and
still have some left over. Maybe not infinite but close enough :slight_smile:

  Tex

Maybe... but only if you can assign all 8-bit characters in domain names.
Last time I checked, it was a significantly small subset of that which were
truly valid.

So you trade dotted quad notation for a base36 number? That defeats the purpose of DNS. Your statement is meaningless, as your local dictionary will show. Randy is right, DNS is finite as well, albeit, constrained semantically, rather than mathematically.

your point?

It would be possible to assign names to every atom in the universe and
still have some left over. Maybe not infinite but close enough :slight_smile:

and that's not half way to even aleph null infinity. :slight_smile:

and if that brings you comfort, you'll love ipv6.

randy

His point was fairly obvious to _me_, Randy: Yes, domain names are not
truly "infinite", but they're much _less_ non-infinite than IP
addresses.

Cheers,
-- jra