small vent

Sun Jun 28 01:17:52 UTC 1998

Brian,

>Keep in mind that some of these "clueless folk" may expect individually
>assigned CIDRs to be registered accurately, and are complaing to the
>registered contact for the block.

APNIC has a policy that _all_ "permanent" reassignments made by ISPs must
be recorded in the APNIC database before address space is considered used.
This policy is recursive; it includes address space assigned to customers
of ISP customers of ISPs (etc), even to the point of individual /32s
assigned in the case of static dialups or virtual hosting.

I believe RIPE-NCC has a similar policy (although their hierarchy is much
shallower that the AP region's for reasons I've never been to clear on).

Of course, the fact that an entry is in a database (any database) doesn't
mean it is valid, either when it was inserted or anytime thereafter, but
that is a generic problem with the registry database "system" (to use the
word loosely), rather than a complaint that can be directed at non-US
registries alone.

>If you know of a more appropriate forum for such a question, I'm
>listening.

More appropriate than NANOG might be "Internet Engineering Planning Group"
<iepg at iepg.org> and "Asia Pacific Operators Forum" <apops at apnic.net>.  I
think RIPE-NCC has something similar (eof at ripe.net?).  You are missing the
point, however.  In any of these forums you are preaching to the converted
-- I'd be surprised (but sadly, not very) if many of the folks that get
mail through iepg, nanog, apops, or the RIPE-NCC counterpart have open
relays.  The people you should be complaining to are the perps themselves
or their upstreams -- they are the only ones who can have an impact (but
you know this).

>Out of curiosity, are you advocating a certain measure of irresponsible
>network management based on sendmail's somewhat cryptic nature?  

I'm advocating nothing.  I provided two possible reasons why it may appear
to you that AP or EU regional networks have more open relays than their NA
counterparts and provided you with the prefixes APNIC allocates to
facilitate your filters.  One could argue that blocking entire continents
is "a certain measure of irresponsible network management", at least from
the perspective of the customer, but I won't argue the point as I assure
you, I am more tired of spam than most.

>God forbid a net manager should open a book.

Ah, sorry.  I must have missed the announcement of Allman's book translated
into Thai, Mongolian, Bangala, (etc.).  Hmm.  Must not have been translated
into English, given the amount of open relays in the US...

Actually, I feel this typifies one of the Internet's Achilles heels --
critical portions of the Internet infrastructure have truly broken defaults
and/or missing/broken documentation.  Information on how to fix those
portions are generally available only through lore or cryptic crib notes
almost always written only in English.  Vendors are typically uninterested
in changing the defaults.  Yet the problems generated by those defaults
have global impact.  Wonder how long it will be before relay is off by
default in Unix distributions (does Sun still have a default broadcast of
all 0s?)...

Regards
-drc