Analyzing different cache replacement policies on cloud

• Main Authors: Khandekar, Apurva A., Mane, Sunil B.
• Format: Conference Proceeding
• Language: English
• Subjects: Algorithm design and analysis; Cloud computing; Computers; Performance evaluation; Servers; System performance
• DOI: 10.1109/IIC.2015.7150834

Today, Caching is considered to be the key technology which bridges the performance gap between memory hierarchies through spatial or temporal localities. Particularly, in disk storage system, it has a prominent effect. To get a higher performance in operating systems, Databases and World Wide Web caching is considered as one of the major steps in system design. In cloud systems, heavy I/O activities are associated with different applications. Due to heavy I/O activities, performance is degrading. If caching is implemented, these applications would be benefited the most. For enhancing the system performance various cache replacement policies have been proposed and implemented and these algorithms defines the enhancement factor and plays a major role in modifying the efficiency of the system. Different caching policies have different effects on the system performance. However, the traditional cache replacement algorithms are not easily applicable to web applications. As the demand for web services is increasing, there is a need to reduce the download time and Internet traffic. To avoid the case of cache saturation and make the caching effective, an informative decision has to be made as to which documents are to be evicted from the cache effectively. This paper gives comparison of different cache replacement policies in traditional system as well as in web applications and proposes a system which implements LRU and CERA caching algorithms and gives it's performance evaluation.