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.
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.
On Mon, Jan 18, 2016 at 10:22 PM, Richard Chycoski ve7cvs@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.
-
Don Fanning -
Richard Chycoski