Database to handle huge amount of data

I'm evaluating a database for my next project. I want to store all the cities in the world (2,5 million) and save weather forecast for every city 开发者_JS百科every day. So you can imagine that the dataset will get quite big fast.

I also need to perform geo queries - get me the city and temperature for this day in this bounding box.

So far I've looked at hbase and couchdb. Hbase looked interesting, but the hardware requirement for production is too expensive for me (a presentation said you need 5 separate servers).

I'd like to keep the costs as low as possible, it's my personal project.

So what other options do I have? Can mongo handle this ammount of data? Anything else?

TL;DR The requirements are

• Large amount of data
• Fast bounding box queries
• Low/cheap hardware requirements
• Optimized for read, but needs to handle insert of 2,5 million records daily

---

Yeah, you can go with mongodb. Mongodb was designed for scaling (sharding, replication). In addition mongodb support geospacial search.