The Entry-Extensible Cuckoo Filter

Stretching your data with taffy filters

“…Additionally, the space usage often grows as $$$ \Omega \left(n\lg n\right) $$$, at which point a traditional hash table would do the same work in the same space with constant‐time operations and an $$$ {n}^{-c} $$$ false positive probability, where $$$ c $$$ depends on the constant in $$$ \Omega \left(n\lg n\right) $$$. A newer approach to manage growing filters is to use cuckoo or quotient filters in which, each time the filter capacity grows, the false positive probability doubles 1,2,19,21,22 . Finally, a third approach to the problem of growing a filter is to depend on the original keys being available during rebuild time 23 .…”

“…A newer approach to manage growing filters is to use cuckoo or quotient filters in which, each time the filter capacity grows, the false positive probability doubles. 1,2,19,21,22 Finally, a third approach to the problem of growing a filter is to depend on the original keys being available during rebuild time. 23 This approach is not always possible or time efficient.…”