Application of both Temporal and Spatial Localities in the Management of Kernel Buffer Cache

Abstract

As the hard disk remains as the mainstream on-line storage device, it continues to be the performance bottleneck of data-intensive applications. One of existing most effective solutions to ameliorate the bottleneck is to use the buffer cache in the OS kernel to achieve two objectives: reduction of direct access of on-disk data and improvement of disk performance. These two objectives can be achieved by applying both temporal locality and spatial locality in the management of the buffer cache. Traditionally only temporal locality is exploited for the purpose, and spatial locality, which refers to the on-disk sequentiality of requested blocks, is largely ignored. As the throughput of access of sequentially-placed disk blocks can be an order of magnitude higher than that of access to randomly-placed blocks, the missing of spatial locality in the buffer management can cause the performance of applications without dominant sequential accesses to be seriously degraded. In the chapter, we introduce a state-of-the-art technique that seamlessly combines these two locality properties embedded in the data access patterns into the management of the kernel buffer cache management. After elaboration on why the spatial locality is needed in addition to the temporal locality, we detail a framework, DULO (DUal LOcality), in which these two properties are taken account of simultaneously. A prototype implementation of DULO in the Linux kernel as well as some experiment results are presented, showing that DULO can significantly increases disk I/O throughput for real-world applications such as Web server, TPC benchmark, file system benchmark, and scientific programs. It reduces their execution times by as much as 53%. We conclude the chapter by identifying and encouraging a new direction for research and practice on the improvement of disk I/O performance, which is to expose more disk-specific data layout and access patterns to the upper-level system software for disk-oriented policies.