Triggers distributed execution process

[Konard]

Triggers are transformations on the transaction log.

That means we can have multiple workers that have a reconstruction of the full links store (database).

Each of the workers can handle only part of triggers for each event, thus making it possible to infinitely parallelize execution of triggers.

After each event it is possible to create a snapshot of all data, that is distributed to all works using transaction log broadcast. After each worker done their part, all parts can be merged back to the transaction log.

This is special case of map-reduce method.