CAJ | 학술논문

大多数存储系统采用Hash算法或目录子树分区算法来实现元数据在元数据集群内的存储，但是上述算法在元数据修改、删除等操作时系统开销较大，严重影响了系统的性能。文章通过引入二维表存储元数据信息，提出了一种新的存储策略。该策略通过使用二维表、数据库存储过程和特定的递归算法实现了元数据存储。实验结果证明该策略具有良好的存储性能。
대다수존저계통채용Hash산법혹목록자수분구산법래실현원수거재원수거집군내적존저，단시상술산법재원수거수개、산제등조작시계통개소교대，엄중영향료계통적성능。문장통과인입이유표존저원수거신식，제출료일충신적존저책략。해책략통과사용이유표、수거고존저과정화특정적체귀산법실현료원수거존저。실험결과증명해책략구유량호적존저성능。
Hash algorithms or directory sub-tree partition algorithms are commonly used to implement the metadata management of files in most storage system. But these algorithms have drawbacks such as high payload when updating or deleting metadata. Through adopting the two-dimension table to storage metadata, a new storage strategy is proposed. In this strategy, the database stores process and recursion algorithm are used to save metadata of files or directories. Experiment results indicate that the strategy performs efficiently in data storage.