A block chain state data storage and reading method based on multi-level caching

Abstract

The invention discloses a block chain state data storage and reading method based on a multi-level cache. in a block chain node, when writing is carried out, according to the minimum unit of a block,all state information in the block and state information obtained after block height coding are written into a MemTable in an atomicity mode, and if writing is successful, the uncoded state information is cached into a first-level cache; And if the MemTable atom writing fails, the state information needs to be written again according to the block number. And if the reading is carried out, readingis carried out according to a sequence of reading the first-level cache and then reading the Memtable by using a conventional method. The method is applied to an existing block chain network, under the condition that it is guaranteed that the block state data can be accurately and completely written in, the storage efficiency of the whole state data is improved, and the method is a new breakthrough in a traditional block chain system based on Key / Value.

Description

technical field [0001] The invention relates to a decentralized blockchain storage system, in particular to a method for storing and reading blockchain state data based on a multi-level cache. Background technique [0002] Blockchain technology, blockchain is a new type of decentralized protocol that can safely store digital currency transactions or other data, information cannot be forged and tampered with, and transaction confirmation on the blockchain is shared by all nodes on the blockchain Completed, the consensus algorithm guarantees its consistency. A public ledger is maintained on the blockchain. The public ledger is visible to any node on the storage block, thus ensuring that it cannot be forged and tampered with. [0003] The storage systems of traditional blockchains are all based on open source Key / Value databases, such as LevelDB. Although LevelDB has greatly optimized the writing and reading of data by using multi-layer merging, the read and write efficiency is...

AI Technical Summary

Problems solved by technology

[0003] The storage systems of traditional blockchains are all based on open source Key / Value databases, such as LevelDB. Although LevelDB has greatly optimized the writing and reading of data by using multi-layer merging, the read and write efficiency is higher than that of other traditional The database has also been improved to a certain extent, but it still cannot achieve the throughput required by actual production for large data volume and continuous block data

[0004] In the traditional blockchain network, all the information in the nodes, including status information and blocks, are stored in the Key / Value database. The block data is huge. Once the system runs for a long time, the block data is too large, which will lead to The storage read and write performance of the block chain network is seriously reduced, which becomes the bottleneck of the operation performance of the entire block chain network and is a major pain point in the practical application of the block chain

Images