On Sun, Feb 19, 2017 at 6:35 AM, Jim Hankins <jim(a)hankinsbay.com> wrote:
This is encouraging. I agree with slide 60, being
new to this, the on boarding has been slow and mysterious. I.e. What specifically is
required in the request, length of time between responses in the request adds a dampening
effect, even if not intended.
In my observation, this is the single greatest problem with AMPR.
Nearly everyone new I have talked to about AMPR reports one of the two
things:
- "I requested a block and never heard back, so I thought the project
was dead. How long does it normally take? It's been a month so far."
- "I requested a /24 for my project linking nnn machines and only got
a /30, so I've had to NAT RFC1918 space and don't really use the AMPR
side for much."
To me, these problems are unacceptable. There are so many subscribers
here who would be willing to join the coordination team and are
competent at assessing block size needs. We need to have multiple
coordinators for each region, so that no coordinator is a single point
of failure.
I'd also like to see some metrics on time-to-allocation, and how often
allocation sizes are different than request sizes (especially
interesting would be comparing this to the utilization of the parent
block--I have a hunch coordinators are being stingy with assignments
when there's no need to be). Once we have some metrics we could set
reasonable goals. Off the cuff, 48 hours to receive an allocation
seems reasonable. I wonder what percent of allocations meet that goal
currently.
If the holdback is technical, I'd be willing to make the required
portal modifications myself. Please send the source code so that I can
get to work.
Tom KD7LXL