HBase good scalability, concurrent write performance is also great.
But HBase actually sacrificed some other features to reach this effect.
Ever since, like some routine operations in MySQL for HBase it might not be so easy.
Let me list a few possible actions.
1. Sort
2. Polymerization
3. nearly a month (annual) statistics, this is really boring, because the data is dynamic, nearly a month (years) this time range is always changing, which is equivalent to a real-time query, this makes me a little headache
4. sort some columns
These requirements are hoping to respond milliseconds or seconds.
You do not know how to solve?