>> Any US east coast devs: does the French mirror get used often for your 
>> archive downloads?
>> 
>> From Boston, MA lil.fr.packages.macports.org responds faster to pings than 
>> packages.macports.org. Since MacPorts uses ping time to calculate where to 
>> make its initial requests from, we might be sending people across the ocean 
>> when it isn't for any actual benefit.
> 
> how do you know it isn't for any actual benefit? (tcp likes less delay...)
> 
>> After spotting the french mirror being used, my initial check saw these 
>> times (ping, 5 samples each):
>> average lil.fr.macports.org response: 102.6ms
>> average packages.macports.org response: 123.8ms
> 
> 
> it would probably be worth comparing file transfer speeds from your location 
> to the two mirrors when you're seeing different ping times like that...

Alrighty, from the numbers: the french mirror transferred data slower but 
responded to pings faster.

$ curl -o /dev/null 
http://packages.macports.org/qt4-mac/qt4-mac-4.7.3_0+quartz+universal.darwin_10.i386-x86_64.tbz2
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  201M  100  201M    0     0   802k      0  0:04:17  0:04:17 --:--:-- 1050k

$ curl -o /dev/null 
http://lil.fr.packages.macports.org/qt4-mac/qt4-mac-4.7.3_0+quartz+universal.darwin_10.i386-x86_64.tbz2
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  201M  100  201M    0     0   559k      0  0:06:09  0:06:09 --:--:--  560k

_______________________________________________
macports-dev mailing list
[email protected]
http://lists.macosforge.org/mailman/listinfo/macports-dev

Reply via email to