BFLUT Bloom Filter for Private Look Up Tables

S Dolev, E Gudes, E Segev,J Ullman,G Weintraub

Cyber Security, Cryptology, and Machine Learning(2022)

引用 0|浏览33
暂无评分
摘要
Open addressing hash tables, possibly under double hashing policy, are regarded more memory efficient than linked list hashing; as the memory used for pointers can be used for a longer table, and allow better-expected performance as the load factor is smaller and there are fewer expected collisions. We suggest further eliminating the single pointer to the memory location used in each entry of the open addressing, and using a single bit per entry, namely use a Bloom Filter to encode the memory address. Thus, obtain even a better load factor, with the same memory, and less number of wrongly mapped addresses when the load is low. Moreover, we can prove that the content in the lookup table that is based on the bloom filter is pseudo-random (in the level of randomization implied by the hash function), thus, keeping the items and the addresses that the LookUp Table (LUT) encodes private.
更多
查看译文
关键词
private look
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要