Hi, if you're asking about the hash table itself, then currently we use linear probing, i.e. k/v pairs with a collision are inserted sequentially starting with the hash%capacity index.