ultra high concurrency queries, hbase

i want to build a mall app , thousands of queries per second for applications used by thousands of users???
does this means that it can not support ultra high concurrency queries(~100000/s)?
i use hbase as database, does it can used as the data source for druid ?

if not?


what is the Druid vs Apache drill?

                                                                                                                         李启明 from China

Hey 李启明,

Druid and Drill are different in that Druid actually stores your data within itself, whereas Drill queries data that lives elsewhere. So if you have an external data source you want to use with Druid, you’d set up an ingestion pipeline to continuously load data into Druid rather than set up Druid to query that data source.

I would take a look at Druid if you need to do “slice and dice” filtered-aggregation kinds of queries where each query would potentially hit a lot of data. These are what Druid is best at.

what about ultra high concurrency queries(100000+/s) ?

在 2016年2月19日星期五 UTC+8上午1:32:03,Gian Merlino写道:

We have had some druid applications that where able to deliver 10K QPS, but it depends on your queries and over all architecture.

can i add mass more nodes in a cluster to achive a almost unlimit QPS(1000k QPS+)?

在 2016年2月21日星期日 UTC+8下午10:54:39,Slim Bouguerra写道:

Cosdyt, we haven’t seen that scale of queries but theoretically it is possible.

Try to read http://static.druid.io/docs/druid.pdf and understand the architecture differences between a key/value store and a column store and their intended use cases.