Struts2 or Django for GAE and future scalability

I am developing location based service. FYI, the database will expand vastly as time and location are the variables. I am considering GAE for initial deployment. I am open for any of python or java based development. While calculating the scalability, I am getting confused. I never thought of scalability before as I haven't worked on big projects. Also I am considering the fact that may be I will have to change hosting in near future for more flexibility.

Considering this situation, what should I start with? Struts2? or Django? Will there be a big difference i开发者_运维知识库n terms of development time?

---

Do you know already know Java or Python? If you are proficient in one and not the other, you might want to use what you know. If you are unfamiliar with both, and particularly programming in general, I think Python would be much easier to learn. But this is very subjective.

GAE is a good platform for some applications. If you are, for example, frequently reporting a location from a mobile device (like a phone), I think GAE would be a good fit. But I would not use django to handle such requests; Instead use the 'lightest' possible framework to record the data (probably webapp (Python) or the low-level datastore API (java)).

Keep in mind the limitations on queries in GAE. There are no JOINS, you'll need to denormalize. You can use inequality filters on one property at a time, so for proximity queries you'll need a technique like GeoBoxes. If you can work around those limitations, App Engine has a lot to offer.