We are developing a simulation-based system for design evaluation of very large distributed file services. This paper describes the workload generation component of our system. We use a parameterized model that extends current models of centralized file system usage, and can reflect the large scale and the sharing patterns that may arise in future distributed systems. The key features of the model are 1) a hierarchical sharing structure, 2) a classification of files according to their sharing and access type properties 3) modeling of client processes and 4) incremental trace generation. 1
Cloud storage systems are currently very popular with many companies offering services, including wo...
File system studies are critical to the accurate configuration, design, and continued evolution of s...
Peer-to-peer systems have quickly become a popu-lar way for file sharing and distribution. In this p...
This paper discusses our general approach to load management in a distributed system, as well as its...
Parallel scientific applications require high-performance I/O support from underlying file systems. ...
Parallel scientific applications require high-performance I/O support from underlying file systems. ...
Most prior studies of distributed file systems have focussed on relatively small communities of user...
The popularity of peer-to-peer file sharing networks has attracted multiple interests even in the re...
In this paper, we describe the collection and analysis of file system traces from a variety of diffe...
This thesis presents a simulation of a distributed file system. It is a simplified version of the di...
Large-scale data-centric systems help organizations store, manipulate, and derive value from large v...
I/O workload is a critical and important factor to analyze I/O pattern and to maximize file system p...
International audienceFile replication is widely used to reduce file transfer times and improve data...
This paper describes the workload characterization of AFS file servers, based on traces collected by...
Abstract: A wide variety of applications such as aerodynamic research, weather forecasting, scientif...
Cloud storage systems are currently very popular with many companies offering services, including wo...
File system studies are critical to the accurate configuration, design, and continued evolution of s...
Peer-to-peer systems have quickly become a popu-lar way for file sharing and distribution. In this p...
This paper discusses our general approach to load management in a distributed system, as well as its...
Parallel scientific applications require high-performance I/O support from underlying file systems. ...
Parallel scientific applications require high-performance I/O support from underlying file systems. ...
Most prior studies of distributed file systems have focussed on relatively small communities of user...
The popularity of peer-to-peer file sharing networks has attracted multiple interests even in the re...
In this paper, we describe the collection and analysis of file system traces from a variety of diffe...
This thesis presents a simulation of a distributed file system. It is a simplified version of the di...
Large-scale data-centric systems help organizations store, manipulate, and derive value from large v...
I/O workload is a critical and important factor to analyze I/O pattern and to maximize file system p...
International audienceFile replication is widely used to reduce file transfer times and improve data...
This paper describes the workload characterization of AFS file servers, based on traces collected by...
Abstract: A wide variety of applications such as aerodynamic research, weather forecasting, scientif...
Cloud storage systems are currently very popular with many companies offering services, including wo...
File system studies are critical to the accurate configuration, design, and continued evolution of s...
Peer-to-peer systems have quickly become a popu-lar way for file sharing and distribution. In this p...