CAJ | 학술논문

为减少软件更新中增量包的大小,提出了基于动态字典的增量更新算法 DICDIFF.该算法使用后缀排序方法预处理新老版本文件,将预处理的结果以后缀数组和名次数组的形式存储为字典目录,基于该字典目录能够快速查找字典数据集与待编码数据之间的相同数据段.随着编码进度的推进,动态扩展字典数据集能够使用更多已知数据段用于构造待编码数据.通过选取多款软件的新旧版本作为实验样本,DIC-DIFF在平均情况下能够节省68．9％的网络流量,高于现有其他增量更新算法.实验结果表明,该算法能够进一步减少增量更新过程中的网络流量.
위감소연건경신중증량포적대소,제출료기우동태자전적증량경신산법 DICDIFF.해산법사용후철배서방법예처리신로판본문건,장예처리적결과이후철수조화명차수조적형식존저위자전목록,기우해자전목록능구쾌속사조자전수거집여대편마수거지간적상동수거단.수착편마진도적추진,동태확전자전수거집능구사용경다이지수거단용우구조대편마수거.통과선취다관연건적신구판본작위실험양본,DIC-DIFF재평균정황하능구절성68．9％적망락류량,고우현유기타증량경신산법.실험결과표명,해산법능구진일보감소증량경신과정중적망락류량.
To reduce the size of delta file for incremental updates,an incremental update algorithm was pro-posed based on dynamic dictionary named differential compression based on dynamic dictionary(DICDIFF). This algorithm uses the suffix sorting algorithm to preprocess the old and new versions.The pretreatments result in the forms of suffix array and rank array which are stored in the dictionary catalogue.Based on this dictionary catalogue,DICDIFF finds the identical segments between the dictionary data set and the data set to be encoded quickly.With the progress of encoding,DICDIFF is expanding the dictionary data set dy-namically.This algorithm gets more segments which could be copied from the dictionary data set to the new version file.Experiments use software from different operating systems as samples,and the average data transmission saving is as high as 69.3%.In comparison with the existing methods,DICDIFF con-structs the minimal delta file.The experimental results show that DICDIFF can reduce the network traffic for incremental updates.