Blockquote
To find places near a location, we have to query all servers and each server will return a set of nearby places. A centralized server will aggregate these results to return them to the user.
Blockquote
If sharding based on region can make some servers hot, sharding based on locationID means every request hits all the servers (and the central server). Doesn’t it make every server even hotter? What’s the point?
1 Like