FAIRIO: An Algorithm for Differentiated I/O Performance

Arunagiri, Sarala; Kwok, Yipkei; Teller, Patricia J.; Portillo, Ricardo; Seelam, Seetharami

doi:10.1109/sbac-pad.2011.26

Cited by 5 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the results show that LFO performs better with large window sizes, using small window sizes allows LFO to be more adaptive to changes in request behaviors. This observation agrees with the common understanding in the literature [21,22,23]. Based on this performance analysis, This paper proposes LFO2, a revised LFO design, which allows LFO to achieve high OHRs using small window sizes.…”

Section: The Effects Of Window Size On Performance and Lfo2supporting

confidence: 88%

See 1 more Smart Citation

LFO2: An Enhanced Version of Learning-From-OPT Caching Algorithm

Kwok¹,

Sullivan²

2021

Machine Learning Techniques and Data Science

Self Cite

View full text Add to dashboard Cite

Recent machine learning-based caching algorithm have shown promise. Among them, Learning-FromOPT (LFO) is the state-of-the-art supervised learning caching algorithm. LFO has a parameter named Window Size, which defines how often the algorithm generates a new machine-learning model. While using a small window size allows the algorithm to be more adaptive to changes in request behaviors, experimenting with LFO revealed that the performance of LFO suffers dramatically with small window sizes. This paper proposes LFO2, an improved LFO algorithm, which achieves high object hit ratios (OHR) with small window sizes. This results show a 9% OHR increase with LFO2. As the next step, the machine-learning parameters will be investigated for tuning opportunities to further enhance performance.

show abstract

Section: The Effects Of Window Size On Performance and Lfo2supporting

confidence: 88%

“…However, it is commonly known that request behaviors, including reuse distance and object size, vary over time [21,22,27]. As a result, a model generated at the beginning of a window may no longer accurately reflect the request behaviors during the later part in the window.…”

Section: Window Size Influences On the Performancementioning

confidence: 99%

LFO2: An Enhanced Version of Learning-From-OPT Caching Algorithm

Kwok¹,

Sullivan²

2021

Machine Learning Techniques and Data Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus, we have been working (both towards the end of the grant period and after) towards the application of this concept of fairness to the design of an I/O scheduling algorithm for a RAID storage system [Ar07,Ar09a]. It will provide best effort guarantees for (a) differentiated service based on the priority of the request class, (b) good device throughput, and (c) latency requirement of request classes.…”

Section: We Designed Two Novel Scheduling Algorithms Cfq-crr (Complementioning

confidence: 99%

DAiSES: Dynamic Adaptivity in Support of Extreme Scale Department of Energy Project No. ER25622 Prime Contract No. DE-FG02-04ER25622 Final Report for September 15, 2004-September 14, 2008

Teller¹

2009

View full text Add to dashboard Cite

The DAiSES project [Te04] was focused on enabling conventional operating systems, in particular, those running on extreme scale systems, to dynamically customize system resource management in order to offer applications the best possible environment in which to execute. Such dynamic adaptation allows operating systems to modify the execution environment in response to changes in workload behavior and system state. The main challenges of this project included determination of what operating system (OS) algorithms, policies, and parameters should be adapted, when to adapt them, and how to adapt them. We addressed these challenges by using a combination of static analysis and runtime monitoring and adaptation to identify a priori profitable targets of adaptation and effective heuristics that can be used to dynamically trigger adaptation. Dynamic monitoring and adaptation of the OS was provided by either kernel modifications or the use of KernInst and Kperfmon [Wm04]. Since Linux, an open source OS, was our target OS, patches submitted by kernel developers and researchers often facilitated kernel modifications. KernInst operates on unmodified commodity operating systems, i.e., Solaris and Linux; it is fine-grained, thus, there were few constraints on how the underlying OS can be modified.Dynamically adaptive functionality of operating systems, both in terms of policies and parameters, is intended to deliver the maximum attainable performance of a computational environment and meet, as best as possible, the needs of high-performance applications running on extreme scale systems, while meeting system constraints. DAiSES research endeavored to reach this goal by developing methodologies for dynamic adaptation of OS parameters and policies to manage stateful and stateless resources [Te06] and pursuing the following two objectives:1. Development of mechanisms to dynamically sense, analyze, and adjust common performance metrics, fluctuating workload situations, and overall system environment conditions.2. Demonstration, via Linux prototypes and experiments, of dynamic self-tuning and selfprovisioning in HPC environments.From a high level, the DAiSES methodology, depicted in Figure 1, includes characterization of application resource usage patterns, identification of candidate (profitable) adaptation targets, determination of feasible adaptation ranges, definition of heuristics to trigger adaptation, design and implementation of OS monitoring, triggering, and adaptation code, and quantification of performance gains.

show abstract