I'm receiving a high speed stream (up to 14 Mbps) of positional information (let's simplify this to (x, y, z, timestamp).

After storing this data, I perform operations such as retrieving points within a bounded 3D space, retrieving points within a timeframe, etc.

I've tried PostgreSQL + PostGIS with sharding by timeframe, but the database can't handle the load. A query for all points within the past hour takes more than 10 minutes.

I have one server (IIRC 8 quad core 2.0GHz CPUs, 32GB RAM). What solutions / technologies / techniques can I use to persist and query the data quickly?

If I had more than one server, what solutions would be ideal?

Thanks!

没有正确的解决方案

其他提示

You may want to have a look at TSDB, if you haven't already done so. It was developed for a use case very similar to what you have described above.

许可以下: CC-BY-SA归因
不隶属于 StackOverflow
scroll top