Hash table processing method

Abstract

The invention provides a Hash table processing method, and a Hash table is an aggregated expandable Hash table and comprises a plurality of expandable Hash tables; each expandable Hash table forms a layer of the aggregated expandable Hash table, each expandable Hash table comprises at least one bucket, and the buckets are used for storing key value pair data; the expandable hash table has buckets with the same number to form a segment of the aggregated expandable hash table; the aggregated scalable hash table includes a shared directory to manage a global depth of the aggregated scalable hash table and pointers of segments. According to the method, the delay abrupt increase and throughput abrupt drop caused by the rehash are effectively reduced, and meanwhile, the high performance of the hash table is ensured.

Description

technical field [0001] The invention relates to the technical field of persistent hashing, in particular to a hash table processing method. Background technique [0002] Dynamic hash tables are widely used in various storage systems, especially memory databases, to achieve efficient query services. With the emergence of a new generation of non-volatile memory (Non-Volatile Memory), some researchers and companies are exploring the porting of hash tables to NVM to implement non-volatile memory databases. NVM is also called Persistent Memory (PM for short). NVM has much higher storage density than DRAM, lower power consumption and persistent capabilities. Although NVM exhibits a write latency similar to that of DRAM and a read latency 3 to 4 times higher, its latency is on the order of hundreds of nanoseconds, which is far lower than traditional persistent storage, such as solid-state drives, disks, etc. Then, building an in-memory database on NVM can not only take advantage...

AI Technical Summary

Problems solved by technology

None of the existing persistent hash tables can achieve the above goals at the same time

Images