Druid vs Batch OLAP?

The biggest challenge that is faced with building a real time OLAP store is being able to ingest a very high volume of data. Druid deals with mostly append-only data and it generates like immutable views over that data. But when you have real time data coming in, it's obviously not immutable because events are constantly being appended on. It requires a little bit more thinking in dealing with a high volume of immutable data versus batches of immutableData.