Competitive prefetching for concurrent sequential I/O

Abstract

During concurrent I/O workloads, sequential access to one I/O stream can be interrupted by accesses to other streams in the system. Frequent switching between multiple sequential I/O streams may severely affect I/O efficiency due to long disk seek and rotational delays of disk-based storage devices. Aggressive prefetching can improve the granularity of sequential data access in such cases, but it comes with a higher risk of retrieving unneeded data. This paper proposes a competitive prefetching strategy that controls the prefetching depth so that the overhead of disk I/O switch and unnecessary prefetching are balanced. The proposed strategy does not require a-priori information on the data access pattern, and achieves at least half the performance (in terms of I/O throughput) of the optimal offline policy. We also provide analysis on the optimality of our competitiveness result and extend the competitiveness result to capture prefetching in the case of random-access workloads.We have implemented the proposed competitive prefetching policy in Linux 2.6.10 and evaluated its performance on both standalone disks and a disk array using a variety of workloads (including two common file utilities, Linux kernel compilation, the TPC-H benchmark, the Apache web server, and index searching). Compared to the original Linux kernel, our competitive prefetching system improves performance by up to 53%. At the same time, it trails the performance of an oracle prefetching strategy by no more than 42%.

Improving OpenStack Swift interaction with the I/O Stack to Enable Software Defined Storage

“…However, such improvements need tools to analyse the whole software stack and the kernel -middleware interactions. The paper by Chuanpeng [13] talks about how to handle concurrent sequential I/O streams in a Virtual Machine setup and explains a similar issue in the VM middleware.…”

Virtual I/O caching

“…Many prefetching designs have emerged to hide I/O latencies [35,36,9,10]. Our work is orthogonal to prefetching but can be used alongside such techniques.…”

“…Recent studies address these issues with various caching techniques. Examples include: (i) adding new metrics for cache replacement policies such as frequency [7], (ii) passing application hints that describe future accesses to the storage server [8], and (iii) identifying future data access patterns and prefetching blocks accordingly [9,10]. These ideas focus on the data that applications use and estimate which data blocks should be cached to improve hit rates.…”

Load balancing regenerating codes for multimedia content streaming

“…Even with a better replication approach, a streaming server handling concurrent access by many users, can still incur in slow disk I/O since the size of the files does not fit on main memory which affects throughput severely. The latter problem is not new and has been studied in the literature before in the form of smarter ways to fetch information from secondary storage [2], [3].…”