Hyperdimensional Hashing: A Robust and Efficient Dynamic Hash Table
TimeThursday, July 14th10:30am - 10:53am PDT
Location3000, Level 3
Design of Cyber-physical Systems, Cloud Computing and IoT
DescriptionMany networking systems rely on hashing algorithms that allow dynamic scaling of the hash table. Examples include AWS, Google Cloud and BitTorrent. Consistent and rendezvous hashing are algorithms that minimize key remapping as the hash table resizes. Neither algorithm offers both efficiency and robustness. Hyperdimensional Computing (HDC) is an emerging computational model that has inherent efficiency and robustness. We propose Hyperdimensional Hashing and show that it has the efficiency to be deployed in large systems. In addition, we show that at a realistic level of memory errors the performance of rendezvous and consistent hashing deteriorates while HD hashing remains unaffected.