Thanks Laura!

Brad, I lived with a Droid for a month and came to a similar conclusion, here's my post: http://arinewman.com/2009/11/droid_almost_does/

I think some will stay and some will go (that pesky 2 year contract). If 30% of the network load dropped off AT&T, presumably the stability would improve and people would be happier as-is. If the stability doesn't improve after the load drops, then we have to point our guns at Apple too.

Some analysts and reports have suggested the phone itself is the problem (a bad chipset, etc). The thing I don't buy about this is that my 1st generation iPhone did not work AT ALL at home in north boulder, but I switched it to TMO (gave it to my wife) and it works great. call drops are non-issue for her. I suppose it could be a combination of an overloaded network/tower infrastructure that can't do call hand-offs, or flawed 3G/EDGE switching but this is not an issue for other devices and other networks.

I also live in NB, and after talking to an AT&T technical support person (L2 or L3) they looked at the distance from the nearest tower and my house, and sure enough AT&T doesn't have a tower any closer than Jay and foothills area, so we are on the far edge of coverage and the downtown tower won't reach well either. I have above ground power lines in front of my house facing east (where the tower is) and basically can't use the phone. Prior to the 850 overlay I got "no service" or 1 bar, now I can see 5 bars but as soon as you pick up the phone and try to use it, it drops out badly. Still not usable. Until they put in a new tower, north boulder is fucked. Ironically, if you are in Dakota Ridge, you are OK...

Next time I'll pay more attention to the set/backdrop. ;-)

I had a pretty frustrating experience with AT&T's recent 850 overlay in Boulder. Right after they did it, my coverage got much better at home, but its since degraded again. Real bummer.

Great comments! I think you've highlighted a key point about how much branding plays a factor, and also the risk/reward of owning the whole stack. It worked against Apple for along time, until it didn't.

Thanks for the write up Frank!

Any cloud developer/service operator should be aware of the inherent limitations of the shared environment, and that runs up and down the stack from hardware to app-layer. The biggest benefit is the instant elasticity of resources, IMO, but clouds come with a lot of issues and you've done a good job highlighting one of them (IP block issues). Are there any real cases where an entire cloud netblock got blocked by a popular web API service for abuse? Would be interesting to see how that worked, why they decided to block at the IP level vs account level, etc.

I'm wondering if this is just moving the problem from one IP to another. If an IP broker is passing API traffic for a bunch of clients, do they run the same risk that their netblock(s) could be blacklisted? Would they have to police usage of their services like email providers do? I like the idea of the service-level authentication/access as the last gate. Twitter actually seems to get this, they are moving from IP whitelists on the search API to account level access on streams.