Hazelcast- in-memory format for distributed computing (static analysis)

Question

In Hazelcast 3.8-EA to run fast-aggregation which is the best in-memory format (OBJECT or BINARY)?

• OBJECT : will fast aggregation api <R> R aggregate(Aggregator<Map.Entry<K, V>, R> aggregator, Predicate<K, V> predicate); skip the serilization/deserilization?
• BINARY : will this add more overhead of serilization? will index work with this format? what is overhead of index?

Answer 1

The indexing will work with both formats.

If you choose the binary format each object will be deserialized in order to extract the attribute specified by the user.

If you choose the object format there will be no deserialization overhead during the extraction.

If you consider the aggregations only it's better to use the OBJECT format.