RAMCloud stores data using log segment structure. When large amount of small files store in RAMCloud, each small file occupies a whole segment, so it may leads to much fragments inside the segments and low memory utilization. In order to solve the small file problem, a strategy based on file classification was proposed to optimize the storage of small files. Firstly, small files were classified into three categories including structural related, logical related and independent files. Before uploading, merging algorithm and grouping algorithm were used to deal with these files respectively. The experiment demonstrates that compared with non-optimized RAMCloud, the proposed strategy can improve memory utilization.
A matching algorithm based on the negative selection for anomaly detection was presented in this paper. In the algorithm the effects of position between two temporal sequence to matching degree were considered. So it could distinguish accurately self and non-self, and reduced the size of detective set effectively. Using normal sequence calls, the initial detective set was created, and the detective set was extended by learning, according to the proportion of anomaly temporal sequence to judge whether this sequence was anomaly. Finally, the results of experiment was given.