Hello, On 06.06.2014, at 12:06, liuguodong wrote:

Our case as following:

1. There are several big tables(more than billion rows per table, more than 1 TB) ; 2. Many dimentions in every table( more than 150 dimentions per table), they random do join operation; 3. Low latency reponse, it is expected to reponse time being less than 1 second.

Could monetdb have a capacity to surpport our case or any suggestion?

This depends first on your queries and second on your hardware. In general, the columnar architecture of MonetDB would be a good match for what you describe (lots of dimensions, random join combinations). I would suggest to simply try your queries with a subset of data on a scaled-down machine (e.g. 100 GB of your data on a 16GB RAM box) and see how it goes. Best, Hannes