[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [Debconf-discuss] cdn.debian.net



Micah and all.
Thank you for reporting CDN.d.n! It is my pleasure.

Unfortunately, CDN.d.n routes according to RTT measurement  and
dynamic BGP information is very difficult.
Because it needs various hosts for measurement like Akamai, Cotendo
CDNetworks and measure CDN players.
And I hope to keep cdn.d.n not to use time consuming web services.

Then, CDN.d.n uses estimation with MaxMind and manual configuration.
Basically, MaxMind distribute lite version of database. It is included
in Debian.
(cdn.debian.net get it by monthly.)


Current impremented:
1) Manually route configuration by CIDR.
2) Manually mirror capability (bandwidth, server response ) configuration.
3) Country estimation according to MaxMind (GeoIP)
4) Continent estimation according to MaxMind (GeoIP also has continent db)

Future plan.
5) BGP AS number estimation according to MaxMind AS.
http://geolite.maxmind.com/download/geoip/database/asnum/
6) latitude and longitude estimation according to MaxMind GeoIP.


Micah,
I think method 5) and 6) supports your request. Please wait it.


On Fri, Aug 13, 2010 at 4:59 AM, micah anderson <micah@riseup.net> wrote:
> On Wed, 11 Aug 2010 12:36:24 -0400, Luca Capello <luca@pca.it> wrote:
>> I guess this discussion should continue on debian-mirrors@, if you drop
>> debconf-discuss@ please Cc: me.
>
> Dropped debconf-discuss, CC'd you.
>
>> I cannot do anything more than thanking Araki-san for this service,
>> which I hope will become official ASAP.
>
> Yes, thank you Araki-san!
>
>> BTW, since we are at it, I am now in Boston (71.232.13.26) and I think
>> cdn.debian.net does not pick up the nearest mirror, which I guess is the
>> MIT one (traceroute hops and mirror FQDN in parenthesis):
>
> I noticed that using cdn.debian.net, what is returned is a RR DNS list
> of ips for mirrors in the U.S. For example, doing an apt-get update I
> got as a response the mirror located in Oregon state university as a
> response. That is almost 3,000 miles away, when Columbia university is
> only .5 miles away.
>
> micah
>



-- 
ARAKI Yasuhiro


Reply to: