【摘要】 文章研究了基于TTL的Web缓存层次过滤效果,业务量性质对基于TTL的动态Web缓存系统的性能有重要影响。在层次缓存中,由于只有错失的请求才会被转发给下一级缓存,因而逐级对业务量存在过滤作用,业务量性质随之改变。文章利用仿真研究了基于TTL的动态Web缓存层次过滤对业务量的影响,重点考察了请求到达间隔模型及对象流行度分布的变化。更多还原

【Abstract】 This paper study the filter effects in TTL-based Web caching hierarchies. The characteristics of workload greatly affect the performance of TTL-based dynamic Web caching system. In caching hierarchies, only the requests that missed in one cache can be forward to the next cache, so each level of cache can consider as one "filter" and changes the characteristics of workload presented to the next level of cache. In this paper, we use simulation to study the filter effects in TTL-based Web caching hierarchies. Experiments focus on the changes of request arrival interval modeling and object popularity distribution of workload. Our work helps to understand the filter effects in caching hierarchies, and implies the guidelines for the design of dynamic Web caching hierarchies and improving system performance.更多还原