19 Jan
2016
19 Jan
'16
12:54 a.m.
Last year I tried to attack the task of cleaning up the DNS. Brian at one
point did an AND against the gateway subnets with the DNS and it produced
~19k A records - hundreds of times what is valid or even being used.
So as a curiosity exercise, I wrote some script to import the DNS into
MySQL. The result was 16653 unique callsigns in the DNS and 2211
aliased/bad call signs as of June 2015.
I then imported the FCC amateur database and compared it to US (A,K,N,W)
callsigns - that produced 3742 unique US callsigns matching the DNS. Out
of those, 1373 callsigns matched an expired, canceled or terminated
license. This doesn't take into account all the vanity call signs for the
1x2/2x2/2x1/1x3 callsigns which may be attributed to another operator.
While we can get good grips on the US callsigns, as soon as it crosses an
international border, it goes to pot as every country has their own
licensing structure. Many do not have online databases. So that's what
we're faced with.
Ideas welcome as to how we can validate the DNS tables.
19 Jan
19 Jan
1:22 a.m.
One of the questions that was asked: Why do we need to validate the
existing tables?
Is the 44 net so congested that having invalid entries in the table a
functional issue?
Are there ways to trim the table? A solution that filters for what are
believed to be unused addresses (based upon utilisation for the last
year or so) might be a good start - the number of false positives should
be small enough for the fallout to be handled with a very small overhead
on the coordinators.
As to the international situation - there are a small number of
countries that have significant allocations that will need to oversee
their own country's IP ranges - but it's important for the largest
'polluter' (the US) to clean up whatever can be recovered.
Is this just a US problem, or are there other ranges where cleanup of
the 44.x space is needed?
- Richard
On 1/18/16 9:54 PM, Don Fanning wrote:
(Please trim inclusions from previous messages)
_______________________________________________
Last year I tried to attack the task of cleaning up the DNS. Brian at one
point did an AND against the gateway subnets with the DNS and it produced
~19k A records - hundreds of times what is valid or even being used.
So as a curiosity exercise, I wrote some script to import the DNS into
MySQL. The result was 16653 unique callsigns in the DNS and 2211
aliased/bad call signs as of June 2015.
I then imported the FCC amateur database and compared it to US (A,K,N,W)
callsigns - that produced 3742 unique US callsigns matching the DNS. Out
of those, 1373 callsigns matched an expired, canceled or terminated
license. This doesn't take into account all the vanity call signs for the
1x2/2x2/2x1/1x3 callsigns which may be attributed to another operator.
While we can get good grips on the US callsigns, as soon as it crosses an
international border, it goes to pot as every country has their own
licensing structure. Many do not have online databases. So that's what
we're faced with.
11:39 a.m.
On Mon, Jan 18, 2016 at 10:22 PM, Richard Chycoski <ve7cvs(a)chycoski.com>
wrote:
(Please trim inclusions from previous messages)
_______________________________________________
One of the questions that was asked: Why do we need to validate the
existing tables?
Is the 44 net so congested that having invalid entries in the table a
functional issue?
Are there ways to trim the table? A solution that filters for what are
believed to be unused addresses (based upon utilisation for the last year
or so) might be a good start - the number of false positives should be
small enough for the fallout to be handled with a very small overhead on
the coordinators.
As to the international situation - there are a small number of countries
that have significant allocations that will need to oversee their own
country's IP ranges - but it's important for the largest 'polluter' (the
US) to clean up whatever can be recovered.
Is this just a US problem, or are there other ranges where cleanup of the
44.x space is needed?
- Richard
It's more about housekeeping and cruft that has built up over the decades
rather than performance. Once you open the floodgates of everyone managing
their DNS records, the task becomes magnitudes more difficult as everything
is then changing. The problem is not isolated to just the US but globally.
As for figuring out what is unused or not, there isn't any good way of
determining that.
3040
days inactive
3040
days old
2 comments
2 participants
participants (2)
-
Don Fanning -
Richard Chycoski