Closed Randoooom closed 1 month ago
Would currently prefer to implement a distibuted cache using skytable. The usage of brokers may be implemented later. The caching dedicated gRPC service does not seem very effective to me tbh, would be hard to gather all remaining service nodes etc. without the implementation of kubernetes drivers etc
closed via #87
In order to further improve the performance on the public (and therefore often used) endpoints we would have to cache the database requests till another endpoints modifies the target dataset.
This is quite easy while working with just one node. With more then one node we would have to sync all the backend caches using events or similar. Possible solutions would be: