> I'm curious about the outcome of this. I don't have enough knowledge > about MongoDB at this point to tell you how well the abstractions in > Storm will work with it, but we can certainly try to help you with > specific issues, and I will personally be curious about the outcome of > your experiment. Even if it turns out to not be a great fit, the > analysis will be worth having.
Well, well, I've been studying storm's code. It would be more dificult than I expected. It seems to me that It's quite impossible to do that wihout changing storm-core, once that storm was built upon SQL way. As I was taking mongoengine code as parameter, I didnt thought storm was so heavy content. I can't imagine a way to just build a backend layer for mongo, in this case we would have to create an SQL--Mongo converter. Another way would be to create a big 'layer' to emulate store-objects, including another almost entire new ORM appended to convert mongo types. A terrible workaround. So the only option would be creating a higher abstraction layer and making current implementation derivate from it, and then derivating the mongo bridge with Storm's ideas and ways. But changing a well-tested core-code is unacceptable. So, does anyone have another idea? Thanks Eduardo Willians -- storm mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/storm
