A request processing method of a hybrid memory key value for a storage system

Abstract

The invention discloses a request processing method for a storage system by using a mixed memory key value, which comprises the following steps of: if a point query request is obtained, searching a hash table and returning a query result; If the skip list is the range query request, constructing a corresponding skip list command, adding the skip list command to a command cache queue of each skip list, waiting until return results of all skip lists are obtained, combining the obtained return results according to the lexical order, and returning the combined return results; If the request is thewrite request, distributing a memory in the NVM and writing a value; Further judging the type of the write request, if the write request is an update request, updating the hash table, and constructing a corresponding table skip command; If the request is the insertion request, newly establishing a skip list node, updating a hash table, and constructing a corresponding skip list command; And determining a target skip list according to the length of the key in the write request, and adding a skip list command to a command cache queue of the target skip list. According to the method, the read-write delay of the medium is fully considered, and the processing performance of the read operation with high proportion is optimized, so that the overall performance of the system is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of information storage, and more particularly relates to a request processing method of a hybrid memory key-value pair storage system. Background technique [0002] Internet big data applications and cloud computing applications require fast access to large-scale data. As a typical representative of non-relational data storage, key-value store (KVS, Key Value Store) adopts unstructured data organization form to provide efficient Data insertion, point query (point query), and range query (scan) support, with its high access performance, high availability and high scalability, can develop rapidly and are widely used in data-intensive applications, such as network index engines, social networks, etc. . Memory key-value pair storage uses Dynamic Random Access Memory (DRAM, Dynamic Random Access Memory) as the storage medium, but DRAM is increasingly unable to meet the capacity requirements of big data applicati...

AI Technical Summary

Problems solved by technology

Studies have shown that the ratio of read data in hybrid memory key-value storage systems is much higher than that of write data, even as high as 30:1, which makes NVM read operation latency a part of the overall system performance that cannot be ignored. Due to the lack of NVM read latency Considering the optimization of system read performance, too many read operations on NVM in the existing index design will lead to a decline in the overall performance of the system

Images