Simulation of client-side caching policies for distributed file systems

Abstract

Cache as an intermediate component serves for storing content which has been requested by user and will be potentially used in the future. The cache capacity is limited, thus we cannot store all requested content. We have to use algorithms which mark an old content to be replaced in the cache. These algorithms are commonly named cache or caching policies. Optimally, the caching policy should replace the content which will not be used in the near future. First goal of this paper is to present a cache simulator which serves for comparison of caching policies. In a simulation, requests for data are produced by random request generator or can be generated from a log file. The second goal of this paper is to present results gained from simulations of different caching policies applicable to mobile devices and other clients in order to choose most suitable one.

Towards a fast and secure design for enterprise‐oriented cloud storage systems

“…Referring to Yamakami, we can conclude that there is pronounced regularity in directory access characteristics for each user. With the increasing frequency of users accessing the cloud storage system, users tend to access the frequently accessed contents . Besides, the directories that have been recently modified and visited previously are more likely to be accessed next time, and a directory with many subdirectories and files also has a high probability of being accessed.…”

“…With the increasing frequency of users accessing the cloud storage system, users tend to access the frequently accessed contents. 23 Besides, the directories that have been recently modified and visited previously are more likely to be accessed next time, and a directory with many subdirectories and files also has a high probability of being accessed. To summarize, the factors that can be used to predict which directories a user may access are directory's size, last accessed time, and accessed frequency.…”