This is a very good question but its also a leading question :)
Datanet provides a level of abstraction such that business logic does not need to be framed in commutative operations rather you just do normal data manipulation on JSON objects and under the covers there is a bunch of commutative operation magic.
CRDTs data guarantee is called strong eventual consistency (SEC), so they are not fit for certain use cases (e.g. take $10 out of accountA and atomically put it in accountB), but the vast majority (~70%) of use cases can be accomplished w/ SEC guarantees (google Peter Bailis' work).
If you can explain a specific business case I can show you how to determine if it is OK to do in SEC
diroussel|9 years ago
Any guidance on how that should be done?
jaksprats|9 years ago
Datanet provides a level of abstraction such that business logic does not need to be framed in commutative operations rather you just do normal data manipulation on JSON objects and under the covers there is a bunch of commutative operation magic.
CRDTs data guarantee is called strong eventual consistency (SEC), so they are not fit for certain use cases (e.g. take $10 out of accountA and atomically put it in accountB), but the vast majority (~70%) of use cases can be accomplished w/ SEC guarantees (google Peter Bailis' work).
If you can explain a specific business case I can show you how to determine if it is OK to do in SEC