The goal of this project is to propose improvements in long-term disk read performance for large-scale high-end storage systems. To accomplish this goal, we first describe access patterns in long-term data. Based on these patterns and current hardware trends, we argue that disk read performance will become increasingly important to overall file system performance. To address this issue, we develop a new layout policy and demonstrate that it improves read performance over other commonly used layout policies. Finally, we discuss trends in high-end storage systems and how the file system-storage system interface can be improved based on the observed access patterns.
I/O performance is a critical aspect of data-intensive scientific computing. We seek to advance the ...
Conventional computer systems have insufficient information about storage device performance charact...
Our objective is to improve disk I/O performance in multi-disk systems supporting multiple concurren...
Disk drives are the bottleneck in the processing of large amounts of data used in almost all common ...
Conventional computer systems have insufficient information about storage device performance charact...
An efficient design for a distributed filesystem originates from a deep understanding of common acce...
Data-set sizes are growing. New techniques are emerging to organize and analyze these data-sets. The...
Benchmarks are important because they provide a means for users and researchers to characterize how ...
Disk drives are the bottleneck in the processing of large amounts of data used in almost all common ...
In this paper, we describe the collection and analysis of file system traces from a variety of diffe...
File system studies are critical to the accurate configuration, design, and continued evolution of s...
The functionality and performance innovations in file systems and storage systems have proceeded lar...
Disk access times have not kept pace with the evolution of disk capacities, CPU speeds and main memo...
In this paper we investigate disk activity caused by accesses to large databases stored on modern st...
Because of the slow access time of disk storage, storage management is crucial to the performance of...
I/O performance is a critical aspect of data-intensive scientific computing. We seek to advance the ...
Conventional computer systems have insufficient information about storage device performance charact...
Our objective is to improve disk I/O performance in multi-disk systems supporting multiple concurren...
Disk drives are the bottleneck in the processing of large amounts of data used in almost all common ...
Conventional computer systems have insufficient information about storage device performance charact...
An efficient design for a distributed filesystem originates from a deep understanding of common acce...
Data-set sizes are growing. New techniques are emerging to organize and analyze these data-sets. The...
Benchmarks are important because they provide a means for users and researchers to characterize how ...
Disk drives are the bottleneck in the processing of large amounts of data used in almost all common ...
In this paper, we describe the collection and analysis of file system traces from a variety of diffe...
File system studies are critical to the accurate configuration, design, and continued evolution of s...
The functionality and performance innovations in file systems and storage systems have proceeded lar...
Disk access times have not kept pace with the evolution of disk capacities, CPU speeds and main memo...
In this paper we investigate disk activity caused by accesses to large databases stored on modern st...
Because of the slow access time of disk storage, storage management is crucial to the performance of...
I/O performance is a critical aspect of data-intensive scientific computing. We seek to advance the ...
Conventional computer systems have insufficient information about storage device performance charact...
Our objective is to improve disk I/O performance in multi-disk systems supporting multiple concurren...