>> 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
