一种闪存敏感的多级缓存管理方法

基于闪存的固态硬盘(solid state driver,简称SSD)已经广泛应用于各种移动设备、PC机和服务器.与磁盘相比,尽管SSD具有数据存取速度高、抗震、低功耗等优良特性,但SSD自身也存在读写不对称、价格昂贵等不利因素,这使得SSD 短期内不会完全取代磁盘.将SSD和磁盘组合构建混合系统,可以发挥不同的硬件特性,提升系统性能.基于 MLC 型 SSD 和 SLC 型 SSD 之间的特性差异,提出了一种闪存敏感的多级缓存管理策略——FAMC.FAMC将SSD用在内存和磁盘之间作扩展缓存,针对数据库系统、文件管理中数据访问的特点,有选择地将内存牺牲页缓存到不同类型的SSD.FAMC同时考虑写请求模式和负载类型对系统性能的影响,设计实现对SSD友好的数据管理策略.此外,FAMC基于不同的数据置换代价提出了适用于SSD的缓冲区管理算法.基于多级缓存存储系统对FAMC的性能进行了评测,实验结果表明,FAMC可以大幅度降低系统响应时间,减少磁盘I/O.

Flash-Aware Multi-Level Cache Management Strategy

Solid state driver (SSD) based on flash memory has been widely used in various types of applications such as mobile devices, PC machines, and servers. Compared with conventional disk, SSD enjoys faster access speed, better shock resistance, and lower power. However, it will not completely replace the disk as the secondary storage in the short run due to its inherent properties such as asymmetric read/write and high price of per gigabyte. Integrating SSD and magnetic disk together can benefit from different performance advantage to obtain good high performance and low cost. This paper proposes a flash-aware multi-level cache scheme (FAMC) which considers the significant discrepancy between MLC type and SLC type SSDs. FAMC uses two types of SSD as a cache layer between main memory and magnetic disk. Depending on the characteristics of data access in database application and file management, FAMC conditionally caches the page evicted by buffer manager to different types of SSD. FAMC considers the impact of the write pattern and type of workloads on the system performance, which adopts flash-aware algorithms and data structures to manage the data stored in SSD. Furthermore, in view of the asymmetry of the replacement cost, the study proposes a flash-friendly buffer replacement policy. The strategy is implemented on a simulation storage system based on SLC type SSDs and MLC type SSDs, and its performance is evaluated. The experimental results show that FAMC can significantly reduce system response time and disk I/O cost.