Richard, You're not lying when you say the resolvers are spitting out different results every minute, now the Cox uplink here goes from Dallas to San Jose to and endpoint in Tokyo. *Insert obligatory Microsoft expletive here* JWP On 8/3/05, Richard A Steenbergen <ras@e-gerbil.net> wrote:
On Wed, Aug 03, 2005 at 01:01:59PM -0500, Justin W. Pauler wrote:
New Zeland and Australia? Me thinks someone goofed. And what's really strange is that Monday I ran this exact same traceroute for informational purposes and at or around hop #7 - cox dallas handed off to atlanta who handed off to msn.net directly.
Seems like the problem has been confirmed to be Savvis. In some places Akadns is returning real Akamai sites with correct routing:
download.microsoft.com is an alias for main.dl.ms.akadns.net. main.dl.ms.akadns.net is an alias for dom.dl.ms.akadns.net. dom.dl.ms.akadns.net is an alias for dl.ms.d4p.net. dl.ms.d4p.net is an alias for dl.ms.georedirector.akadns.net. dl.ms.georedirector.akadns.net is an alias for a767.ms.akamai.net.
In others it is returning download.microsoft.com.c.footprint.net:
download.microsoft.com is an alias for main.dl.ms.akadns.net. main.dl.ms.akadns.net is an alias for dom.dl.ms.akadns.net. dom.dl.ms.akadns.net is an alias for download.microsoft.com.c.footprint.net
The footprint.net CDN appears to be updating every minute or so with new data, each more broken than the last. So far this morning I've also seen (from various places not near any of the final sites):
cdn-colo.Frankfurtfrx.savvis.net 90.626 ms ge-0-0-1.a20.taiptw01.tw.ra.verio.net 197.312 ms so-2-0-1-0.par22.ip.tiscali.net 82.875 ms iadvantage-1.gw2.hkg3.asianetcom.net 238.817 ms xe-0-1-0.a20.osakjp01.jp.ra.verio.net 271.745 ms bcr1-so-1-0-0.Londonlnx.savvis.net 167.918 ms cpr2-pos-0-0.VirginiaEquinix.savvis.net 90.493 ms ae-11-51.car1.Atlanta1.Level3.net 149.736 ms
If you really want to download something from microsoft at faster than crawl speeds, try moving your resolvers around to find a nameserver that is getting real Akamai results. Otherwise, wait for Savvis and/or MS to get their act together. :)
-- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/ras GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)