Some of the value I see from projects like Simple Cloud, would be to help alleviate the fear of vendor lock in.

While this is not as big a deal as you have pointed out, it’s nice to have the tools that you can take to a decision maker to allow you to more easily get buy in.

In the end, projects like this may not be of value to everyone, but they will drive the cloud forward and I think that’s a positive thing.

1) The DB abstraction analogy is very appropriate in many ways, but IMO you take it pretty far without tying it back to the API. Yes, there will be what we call ‘vendor value add’ features that more advanced cloud applications are likely to use. We’ve openly embraced this fact. We offer access to the basic functionality with the Simple Cloud API, and we have the luxury of having full-featured client libs that developers can drop down to in ZF. So, why use the API if you plan to use such features? For one, you will have a much clearer distinction between value-add features and common features so that you know when you’re locking in and can more easily migrate if you must. Also, you will also have only one very simple and stable API to learn, even if you switch vendors. The Simple Cloud API is also *extremely* thin to the point that it adds negligible overhead to the calls made by the client libs in ZF. There is no magic under the hood, and no generation of anything, so I’m not sure what the complex SQL generation has to do with anything here to be honest.

2) See 1)

3) Keep in mind that Amazon hasn’t said no yet. In fact, we have an ongoing discussion with them, and they are very interested in what we’re doing both in the API and the AWS client libs in ZF. But regardless of whether they contribute to the Simple Cloud API, the adapters for their services are still full featured due to community and Zend’s own contributions. While Amazon will always be welcome to join the effort- along with any other cloud player that can contribute- we recognize that they are the leader in the cloud space and have therefore ensured we support them well. I also personally believe that MS, Rackspace, Nirvanix, and GoGrid are true leaders in their cloud offerings, and IBM is leading thought around cloud technologies. I also think IBM will have a great impact and influence on the cloud space if/when they release their own offerings.

4) Well, adoption of the Simple Cloud API isn’t hard either. And there are still benefits even if you don’t switch (see 1)).

5) True, established apps may never switch. But developers evaluating services for their barely written apps (likely nothing more than prototypes) are likely to switch around until they find the right platform for them. After they’ve found the right service they’ll see no reason to move off of the Simple Cloud API if we’ve met our goals with the project. And keep in mind that not all cloud apps have the cloud requirements that your apps have.

In any case, thanks for the feedback. I’d really appreciate it if you started a forum topic about this here: http://www.simplecloud.org/forum. And/or dicuss this with me in an open forum like a webinar, etc.

Thanks again!
,Wil