MongoDB: A first look | Hacker News Reader

MongoDB: A first look

travisswicegood.com

27 pointstswicegood7 comments

It seems that a lot of its users are the former LAMP crowd. I've been using it as a distributed file server for configuration type files. Works great. The APIs are really simple and feel native to the language and it has all the basic datatypes you'd want for configuration.

http://www.businessinsider.com/ uses MongoDB http://www.businessinsider.com/how-we-use-mongodb-2009-11

One weekpoint of MongoDB that's worth noting is that it's dependent on little endian architecture. I was going to useI was going to use MongoDB in a project not long ago until I discovered this. The project would have been required to run on SPARC.

I have see a lot of MongoDB articles lately. I am curious if anyone here has used it or played with it and what your thoughts about it are?

I use it in production on http://www.ocwsearch.com/ .

Short answer: it's awesome. Key points:

1. It's fast. Haven't benchmarked it properly yet, but with decent traffic and a decent volume of stored data, it's not showing no signs of problems. Make sure your indexes are correct.

2. Very helpful community. The users mailing list is filled with helpful users and thr 10gen people (who make mongo) are active there too.

3. Very easy to set up and use from PHP. The driver just works after a simple installation.

There are lots of docs to get you started- Google for tutorials with your favorite language. Also, 10gen run conferences of mongo users and the presentations are usually excellent. I'm giving a talk at the London conference on 18 June and I think their is a Paris one too soon.

I'm using it for a handful of internal projects. The selling points for these scenarios were that the server is cross platform, the server footprint is really small, and there is rich and mature-enough library support to use it from just about anywhere.

It's main purpose is to act as a quasi-cache for a bunch of database queries. A couple of batch apps pull chunks of data from a bunch of databases, essentially consolidating and precalculating a bunch of stuff.

The biggest strength we've found so far is that it is schema-less. We're not spending much time thinking about how to organize the data, just getting it and sifting it.

Hope this helps

We're using it for one of our subsystems (user history) at bit.ly. We've been quite pleased with the performance and flexibility.

Reply on news.ycombinator.com