Simultaneous Evaluation of Multiple I/O Strategies

2016

The Computer Journal

Self Cite

This paper presents two mechanisms that can significantly improve the I/O performance of both hard and solid-state drives for read operations: KDSim and REDCAP. KDSim is an in-kernel disk simulator that provides a framework for simultaneously simulating the performance obtained by different I/O system mechanisms and algorithms, and for dynamically turning them on and off, or selecting between different options or policies, to improve the overall system performance. REDCAP is a RAM-based disk cache that effectively enlarges the built-in cache present in disk drives. By using KDSim, this cache is dynamically activated/deactivated according to the throughput achieved. Results show that, by using KDSim and REDCAP together, a system can improve its I/O performance up to 88% for workloads with some spatial locality on both hard and solid-state drives, while achieves the same performance as a "regular system" for workloads with random or sequential access patterns.

Section: Experiments and Methodologysupporting

confidence: 86%

Section: Aged File Systemsupporting

confidence: 60%

See 1 more Smart Citation

Improving I/O Performance Through an In-Kernel Disk Simulator

González-Férez

Piernas

2016

The Computer Journal

Self Cite

“…Our approach is limited to non-critical systems because of the learning periods. Methods using online simulation as [17] need to build the disk simulator (virtual disk) and modify the I/O schedulers. They do not need learning periods, but their capabilities are reduced as device complexity increases.…”

Section: Related Workmentioning

confidence: 99%

Automatic I/O Scheduler Selection through Online Workload Analysis

Nou

Giralt

2012 9th International Conference on Ubiquitous Intelligence and Computing and 9th International Conference on Autonomic and Tr

2012

Self Cite

I/O performance is a bottleneck for many workloads. The I/O scheduler plays an important role in it. It is typically configured once by the administrator and there is no selection that suits the system at every time. Every I/O scheduler has a different behavior depending on the workload and the device. We present a method to select automatically the most suitable I/O scheduler for the ongoing workload. This selection is done online, using a workload analysis method with small I/O traces, finding common I/O patterns. Our dynamic mechanism adapts automatically to one of the best schedulers, sometimes achieving improvements on I/O performance for heterogeneous workloads beyond those of any fixed configuration (up to 5%). This technique works with any application and device type (RAID, HDD, SSD), as long as we have a system parameter to tune. It does not need disk simulations or hardware models, which are normally unavailable. We evaluate it in different setups, and with different benchmarks.

“…Most of the times, they do not make any selection, and the default scheduler (CFQ) is used, when a different one could improve the throughput. Therefore, we present the design and implementation of a Dynamic and Automatic Disk Scheduling framework (DADS ) that, by using an enhanced version of an existing in-kernel disk simulator [2], is able to automatically and dynamically select the best Linux I/O scheduler by comparing the expected performance achieved by each one. Our first implementation compares two schedulers, by running two instances of the disk simulator.…”

Section: Introductionmentioning

confidence: 99%

Dads

González-Férez

Piernas

Proceedings of the 27th Annual ACM Symposium on Applied Computing

2012

Self Cite

The selection of the right I/O scheduler for a given workload can greatly improve the performance of a system. But, this is not an easy task because several factors should be considered, and even the scheduler deemed the "best" can change at any moment. So, we present a Dynamic and Automatic Disk Scheduling framework (DADS ) that compares different Linux I/O schedulers and automatically and dynamically selects that which achieves the best performance for any workload. The implementation described here compares two schedulers by running two instances of a disk simulator inside the Linux kernel, each one having a different scheduler. Our proposal compares the schedulers' service times, and changes the scheduler in the real disk if the performance is expected to improve. DADS has been analyzed by using different workloads, hard disks, and schedulers. Results show that it selects the best scheduler of the two compared at each moment, improving the performance and exempting system administrators from selecting a suboptimal scheduler.