I'm glad we can put the "upload cap" theory to rest. However, some parts of AT&T's response do not make sense and are just plain inaccurate:
"...we are providing normal 3G uplink speeds and consistent performance for affected customers with HSUPA-capable devices."
100 kbps is NOT "normal 3G uplink speeds", though I'll give them the "consistent" part. It definitely has been consistently near 100 kbps for the last 5 days! But "normal" non-HSUPA upload speed is certainly faster than 100 - more like 200-300 with a theoretical limit of 384 kbps.
Wired is even reporting that a temporary "fix" is in place, but I would beg to differ:
"AT&T says a temporary fix is now in place, so upload speeds for AT&T 3G users should be back to normal...."
http://www.wired.com/epicenter/2010/07/att-blames-slow-3g-on-alcatel-lucent-bug/#ixzz0t1RUAI63
AT&T also says that this impacts
"...less than two percent of our wireless customer base." But it's much larger than that if you count the non-HSUPA devices that have also been affected by the 100 kbps limit. For example, I have an iPad 3G that also shows speed tests of 100 kbps uploads, when it normally would be in the 200-300 range. Granted, that's not as big of a difference as 1600 to 100, but still noticeable.
Also, how is a "software bug" responsible for suddenly dropping upload speeds from 1600 to 100 overnight? Did they perform some kind of software "upgrade" in these markets that went wrong? Was there some kind of Y2K bug that just appeared on July 1? They don't say, but I must admit this still seems a bit suspicious.
And finally, they have no idea when the fix will be in place. That's just great! Could be a day, could be a week, could be a month! They run one of the largest networks in the world and many of their major metro areas are experiencing this issue, and yet they can't even pinpoint a timeline?