s3u

Right. One of ql.io's goals is to enable (b) but on a tier between clients and servers.

OData specifies a resource model, resource metadata and URI naming conventions for HTTP APIs. The query formats in OData are part of the producer-implemented APIs as queries happens at the origin. ql.io is a layer above such APIs, and generally operates across several such APIs.

Thanks for the "hint". Checking.

Exactly. We tried to stay close to the SQL syntax to a large extent although there is no relational storage underneath.

Any reason your tests used HTTP/1.0? It is likely that most of the time is spent in connection establishment and you may not have tested the server.

The source is now on github. See the links at the top of the blog post.

The source is now on github. See the blog post for github links.

What's misleading about them?

Please feel free to reproduce on your own. The source is on github. See the blog post for github links.

Yes, correct - I'm not able to disclose the code. But there is more to come and I will be sharing the code on github.

Point well taken - there was no such intention. I updated the charts to show req/sec and mean response time separately. See http://www.subbu.org/blog/2011/03/nodejs-vs-play-for-front-e...