Your browser doesn't support javascript.
loading
Show: 20 | 50 | 100
Results 1 - 3 de 3
Filter
Add filters








Language
Year range
1.
Journal of Pharmaceutical Analysis ; (6): 9-13,23, 2008.
Article in Chinese | WPRIM | ID: wpr-624698

ABSTRACT

At present, there are many effective ways to achieve high performance in cluster system storage management, including server-end disk, server-end caching, local caching and cooperative caching. The cooperative caching mechanism shares caches among different clients so as to avoid expensive disk access costs and to improve overall throughput of cluster system. In this paper, a Single Copy Cooperative Cache model is proposed together with block lookup algorithm, block replacement algorithm and the consistency algorithm based on the model. Meanwhile, the prototype system of the model is implemented in PVFS file system. Finally, the performance of this system is tested in InfiniBand Framework, the result of which shows that in contrast to the original PVFS system, read performance of PVFS file system is improved by about two times, while write performance is reduced by nearly ten percent.

2.
Journal of Pharmaceutical Analysis ; (6): 51-55, 2007.
Article in Chinese | WPRIM | ID: wpr-621737

ABSTRACT

Influences of the clock resolution of bandwidth estimator on the accuracy and stability of the packet pair algorithm was analyzed. A mathematic model has been established to reveal the relationship between the result deviation coefficient and the packet size, clock resolution and real bandwidth(value)of the measured route. A bandwidth self-adapting packet pair algorithm was presented based on the mathematic model to reduce the estimation error resulting from the clock resolution and to improve the accuracy and stability of measurement by adjusting the deviation coefficient. Experimental results have verified the validity and stability of the algorithm.

3.
Journal of Pharmaceutical Analysis ; (6): 131-135, 2006.
Article in Chinese | WPRIM | ID: wpr-621755

ABSTRACT

Objective Network emulation system constructs a virtual network environment which has the characteristics of controllable and repeatable network conditions. This makes it possible to predict the availability and performance of new protocols and algorithms before deploying to Internet. Methods PARNEM, a parallel discrete event network emulation system described in this paper has the following characteristics: ① BREEN - a BSP based real-time event scheduling engine; ② application transparent flexible interactive mechanism; ③ legacy network model reuse. Conclusion PARNEM allows detailed and accurate study of application behavior. Comprehensive case studies covering bottleneck bandwidth measurement and distributed cooperative web caching system demonstrate that network emulation technology opens a wide range of new opportunities for examining the behavior of applications.

SELECTION OF CITATIONS
SEARCH DETAIL