Order Tray | Contact Us | Home | SIG Lists

[aprssig] Tier 2 Update

dick at kb7zva.com dick at kb7zva.com
Thu Jun 22 20:27:29 UTC 2006


It seems strange to me that we have to take the time to explain something so obvious.
The advantage of spreading out the load and bandwidth should be a no-brain-er.

If you ask any network tech if he wanted all his servers at a few locations
state-side or to spread them out world-wide, I'm sure that about 99% would say
that it makes more sense to spread them out. That way if you have a melt down
at one location it doesn't stop everything. And, after the last Hurricane season
I'm sure there are a lot of people down south wishing they had spread out some
services the same way.

It's probable that we won't ever see all core servers disappear at the same time.
That doesn't mean we won't lose hubs across the pond into EU, or for that matter
hubs in the USA. The wider we can spread our T2 service the more effective it
becomes and avoids larger single_points_of_failure.

We've tested many possible contingency plans in the event of outages, including
the possibility of what might happen if the core wasn't available. Losing one
low yield T2 server isn't as dramatic losing a core server that has 200 or
more clients. Everything that points to wide-spread distribution is more 
effective.

Tier 2 has grown to 31 regular servers, 4 WX servers, and about 6 specialized
servers. Currently we are testing 4 more servers soon to come on line. Should
we be concerned about connectivity for 45 servers? I think so. All I need to
know is which of the core sysops or leader has the authority to make decisions
or offer support for the entire aprs.net? So far that isn't clear to us.

As Steve said, 'those who write code, rule' (something like that).
If he wasn't joking when does Pete get the keys? <g>

Dick, KB7ZVA
APRSWest  

  





More information about the aprssig mailing list