Notes on snappydata

• A new IMDG focus on OLTP + OLAP, streaming
• the core idea is they use Spark + Gemfire
• Spark for OLAP solutions
• Gemfire for OLTP solutions
• When data be imported, it can be built on row/column based tables
• how did they do updates for column based data?
• they use delta row update (merge updates in memory) and do lazy update to column based data
• each data will be stored in memory, and can overflow to disk when memory is tight

Sincerely, Yuan