You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, in order to load a table entity, there will be at least four IO accesses (three for id-name mappings and one for real load), which is very time-consuming and costly.
Considering there will be only a few metalake and catalog entities and those entities are almost unchanged, So we propose we can cache the id-name mapping of those entities to reduce the access time.
How should we improve?
We can use Guava cache or Caffeine cache to keep the ids that map to metalakes and catalogs in memory.
Note: We should be careful about the consistency issue between the backend storage and cache
The text was updated successfully, but these errors were encountered:
What would you like to be improved?
Currently, in order to load a table entity, there will be at least four IO accesses (three for id-name mappings and one for real load), which is very time-consuming and costly.
Considering there will be only a few metalake and catalog entities and those entities are almost unchanged, So we propose we can cache the id-name mapping of those entities to reduce the access time.
How should we improve?
We can use Guava cache or Caffeine cache to keep the ids that map to metalakes and catalogs in memory.
Note: We should be careful about the consistency issue between the backend storage and cache
The text was updated successfully, but these errors were encountered: