Page scheduling algorithm based on periodic access and latency cost
1.School of Computer Science and Engineering, South China University of Technology, Guangzhou, Guangdong 510641, China;
2.Department of Computer Science and Technology, Guangdong University of Finance, Guangzhou, Guangdong 510521, China;
3.Science Research Department, Guangdong Polytechnic Normal University, Guangzhou, Guangdong 510665, China)
Abstract: Due to the difficulty for existing page scheduling algorithm to further improve system performance by increasing hit rate, a new algorithm was proposed. The information of page access time, location and counts was statistically collected to analyze the time scope of reading latency, and a cache buffer management data structure with a cost bit was designed. Based on periodic access pattern and latency cost, a cache buffer replacement algorithm was proposed with scheduling simulation and practical implementation. The results show that the page access pattern shows periodicity for expanding observing time scale with crossinstruction cycle. By swapping out the pages with high latency cost as later as possible to avoid cost cyclic access, the periodicity plays an important role in page scheduling for shortening the average queue length and waiting time of storage system. The system performance can be improved by the proposed page scheduling algorithm.