A quantitative analysis of cache policies for scalable network file systems
SIGMETRICS '94 Proceedings of the 1994 ACM SIGMETRICS conference on Measurement and modeling of computer systems
Cello: a disk scheduling framework for next generation operating systems
SIGMETRICS '98/PERFORMANCE '98 Proceedings of the 1998 ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems
Capacity planning with phased workloads
Proceedings of the 1st international workshop on Software and performance
Bandwidth allocation in a self-managing multimedia file server
MULTIMEDIA '01 Proceedings of the ninth ACM international conference on Multimedia
Managing energy and server resources in hosting centers
SOSP '01 Proceedings of the eighteenth ACM symposium on Operating systems principles
Minerva: An automated resource provisioning tool for large-scale storage systems
ACM Transactions on Computer Systems (TOCS)
Performance Guarantees for Web Server End-Systems: A Control-Theoretical Approach
IEEE Transactions on Parallel and Distributed Systems
Introduction to Reinforcement Learning
Introduction to Reinforcement Learning
Aqueduct: Online Data Migration with Performance Guarantees
FAST '02 Proceedings of the Conference on File and Storage Technologies
Appia: Automatic Storage Area Network Fabric Design
FAST '02 Proceedings of the Conference on File and Storage Technologies
Selecting RAID Levels for Disk Arrays
FAST '02 Proceedings of the Conference on File and Storage Technologies
Hippodrome: Running Circles Around Storage Administration
FAST '02 Proceedings of the Conference on File and Storage Technologies
ISTORE: Introspective Storage for Data-Intensive Network Services
HOTOS '99 Proceedings of the The Seventh Workshop on Hot Topics in Operating Systems
Feedback Based Dynamic Proportion Allocation for Disk I/O
Feedback Based Dynamic Proportion Allocation for Disk I/O
Recovery Oriented Computing (ROC): Motivation, Definition, Techniques,
Recovery Oriented Computing (ROC): Motivation, Definition, Techniques,
Scalable content-aware request distribution in cluster-based networks servers
ATEC '00 Proceedings of the annual conference on USENIX Annual Technical Conference
Triage: Performance differentiation for storage systems using adaptive control
ACM Transactions on Storage (TOS)
Controllable fair queuing for meeting performance goals
Performance Evaluation - Performance 2005
eQoS: Provisioning of Client-Perceived End-to-End QoS Guarantees in Web Servers
IEEE Transactions on Computers
Non-intrusive performance management for computer services
Proceedings of the ACM/IFIP/USENIX 2006 International Conference on Middleware
Pricing-based strategies for autonomic control of web servers for time-varying request arrivals
Engineering Applications of Artificial Intelligence
Differentiated storage services
ACM SIGOPS Operating Systems Review
DynaQoS: model-free self-tuning fuzzy control of virtualized resources for QoS provisioning
Proceedings of the Nineteenth International Workshop on Quality of Service
Differentiated storage services
SOSP '11 Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles
Non-intrusive performance management for computer services
Middleware'06 Proceedings of the 7th ACM/IFIP/USENIX international conference on Middleware
A self-tuning fuzzy control approach for end-to-end QoS guarantees in web servers
IWQoS'05 Proceedings of the 13th international conference on Quality of Service
Hi-index | 0.00 |
In this paper, we address the problem of dynamic allocation of storage bandwidth to application classes so as to meet their response time requirements. We present an approach based on reinforcement learning to address this problem. We argue that a simple learning-based approach may not be practical since it incurs significant memory and search space overheads. To address this issue, we use application-specific knowledge to design an efficient, practical learning-based technique for dynamic storage bandwidth allocation. Our approach can react to dynamically changing workloads, provide isolation to application classes and is stable under overload. We implement our techniques into the Linux kernel and evaluate it using prototype experimentation and trace-driven simulations. Our results show that (i) the use of learning enables the storage system to reduce the number of QoS violations by a factor of 2.1 and (ii) the implementation overheads of employing such techniques in operating system kernels is small.