Hi all, as I read here https://cwiki.apache.org/confluence/display/KYLIN/KIP-1%3A+Parquet+storage the new storage engine will be parquet and the queries will go through apache spark jobs. I'm not so familiar with apache spark but I think that it's not easy to tune it to support a lot of concurrent queries. What would be the guideline for configuring kylin + parquet to support for example 100 concurrent queries per second? What kind of deployment (machine types) and configurations will work in that case?
-- Greetings, Ivan Georgiev
