This looks interesting, however I have several reservations about this. First and formost is the pricing issue. In the recent past Google has not been very smart about pricing it's cloud offering and the Google Cloud SQL appears to be making a similar mistake.

The official announcement says that it's free for now but if/when they change that you will be notified in advance. I know that sounds good but to me it screams vendor lock-in. My thought is that 6 months from now Google will want $6 / second or something ridiculous and I'll be scrambling to move platforms.

The other thing is it raises the spectre of the future of MySQL in Oracles hands. Who knows how much longer MySQL will be viable?

Both pieces of information make me feel like I would be building a system on a pile of sand that can be swept away in an instant. That's not a comfortable feeling...

Wait, what would you suggest they do instead of announcing the pricing 30 days in advance? That seems completely reasonable to me... Also, considering several huge sites still use MySQL like Facebook, Wikipedia, etc, I think you'll be fine in that respect for the foreseeable future.

Why not determine pricing before they launched into beta or at least provide guidance numbers? The problem with the 30 day warning in pricing is that the exact same thing that happened with gae pricing could happen here: someone builds a project with a certain price range in mind, pricing is announced and is far higher than expected for this use case, developer has 30 days to scramble to migrate their app and write angry blog posts.

Starting out as free is a mistake in my mind, because it doesn't properly set developer expectations from the start about what those price levels will be or even what metrics the price will be based upon. I would bet that no matter what they do, there will be some people who will be upset when pricing is finally announced.

Well, I think for precisely those reasons, people will know not to form their own pricing expectations in their mind this time.

I hope they add support to PostgreSQL.

I believe Google releases these things for free so that people can help them test it out and help them find the right pricing model. I'm not too sure there are a lot of people that will immediately base their business model on a service that has not yet announced the pricing. I think Google has been extremely reasonable with App Engine. Also, Lock-in can be relatively avoided if you write your code correctly.

This is a huge step forward, can't wait to dive into this for CoderBuddy.

I do agree with others though that there should be more than 30 days notice on pricing. And it would be nice to have some guidance on pricing, given recent history...

Why MySQL over PostgreSQL? I would like to know how they choose one over the other...

I don't know their official reason, but they seem to have a lot of in-house knowledge of mysql internals [1].

1: http://code.google.com/p/google-mysql-tools/

From what I understand they use or at least used to use Mysql internally. So they have experience scaling Mysql.

Maybe the app engine team just chose the one they had more experience in hacking.

We're very very interested in this, but given the previous bait and switch on pricing, we're not going to touch it with a 10 foot pole until the pricing is set.

from working with gae excessively over a year now, gae is not really reliable enough for mission critical systems. Datastore queries that work one day, might not work the next, random exceptions using the gae api is common.

so we can sql with our sql while we sql?