20 Feb
2017
20 Feb
'17
6:04 a.m.
On Sun, Feb 19, 2017 at 2:41 PM, Brian Kantor <Brian(a)ucsd.edu> wrote:
Thanks Tom. The portal already automatically sends
reminder notices
to coordinators with outstanding allocation requests, twice a month.
If all the coordinators agree, we could change that to weekly.
I find that the coordinations which take more than a few minutes to handle
are those where the requester is confused and asking for something that
is not sensible, so an email exchange is necessary to get the request
straightened out. And it takes time to verify that the requester has
supplied a valid callsign that matches his name. You might be surprised
at how many folks just want a block of addresses and make up a callsign
to try to get some. We've even had cases where the person registered
a friend of his who is a ham but who has no interest in AMPRNet; they
just were stealing his identity to get a block of addresses.
And there are the BGP announced allocations, which require them to fill in
an additional form by email. (It's at this point that I often find out
that the person has no idea what BGP is and just checked the box out of
whim, or hasn't asked his ISP if they'll announce a block for him, etc.)
As for me granting coordinations, for historical reasons, each coordinator
may have his own way of allocating subnets. Since I can't know all
those schemes, it's not practical for me to just take a stab at it.
This is a volunteer organization, and people devote their time to it
as their schedule permits. I think the great majority of the volunteer
coordinators are doing a good job, and I don't think the system is broken.
When a coordinator seems to be unresponsive, I contact them to ask if I
can help. On these rare occasions, the allocation request just slipped
into a crack and a simple reminder is all that's necessary. If they
don't respond, I find another coordinator by looking for volunteers.
- Brian
I agree that most of the time the system works adequately. I
absolutely believe there's room for improvement.
Automatic systems can be used to verify callsigns, eliminating much of
the latency in legitimate requests. Multiple schemes could be used
including using HTTPS with Logbook of the World certificate based
client verification (we use this method for authentication within the
HamWAN network, both with and without encryption). Databases like FCC
ULS could be automatically cross-referenced in cases where they didn't
present a certificate. When automatic methods are exhausted, the
request can be dropped into a queue for manual verification. This
verification queue can be handled by any responsible volunteer,
freeing up the coordinators' time to focus on actual allocations.
For BGP, application form labeling and fields can be improved so there
are fewer accidental selections. By adding the NSP fields to the
application, you could save the time it takes to ask for more
information and generate your form letter.
To support assistant coordinators (eliminating single point of
failure), there could be a section for the coordinator to document
their allocation plan logic. If documented properly, anyone could
allocate space to the coordinator's specification without the
coordinator having to take any action. This also provides an
opportunity for peer review in cases of misguided allocation schemes
(such as breaking up a state block by county).
I appreciate the time you've taken to enumerate these challenges so
that we can discuss ways to improve them. I'd love to help improve the
AMPR systems.
Tom