Hello.
Main difference:
- PrestoDB is a query engine for Hadoop, it uses HDFS for storage;
- ClickHouse uses its own storage.
If you have just execution engine, then to process interactive queries efficiently, you need to prepare your data in advance.
For example, you must collect data for each day and then sort the data by primary key, to help processing of range queries.
In contrast, ClickHouse allows to maintain data locality by primary key, while continuously ingesting new data.
You could insert incoming data in near-realtime and to process analytical queries without degradation of performance.
Also you could find some user experience here:
https://groups.google.com/d/msg/clickhouse/pwow5kIj7O0/87phXepFBwAJ