hank
hank copied to clipboard
Add ability to query by hashed key directly
This would allow us to not compute the hash value multiple times and more importantly we could avoid the cost of transferring entire keys over the network if they are large.
There are some tradeoffs, though - for instance, it will be impossible to tell what actual keys are being queried from the server's perspective.
True. I imagine we could give the user the option to choose between querying by key or querying by hashed key and have clients support both. The former is easier to use (no need to deal with a Hasher) and easier to debug (the server knows what keys are being queried), the latter is more performant -- especially if keys are large.