Performance evaluation of our proposed asynchronous data replication in the distributed system. will be reported. The distributed system is composed of one primary copy server and several replica servers. Our proposed architecture features an, asynchronous data replication by time division, of the database update. In other words, the database allocated to each replica server is respectively updated by arriving transactions, flushed to the primary copy and propagated to every replica server in asynchronously time-division manner. Our replication algorithm is designed using the log as record of the updated information and two kinds of queues at each replica server. Experimantal results of our asynchronous data replication is obtained to evaluate appropriate synchronized period in asynchronous data replication for the purpose of load reduction at each server. The experimental data are (1) time series of CPU occupation rate under combination of transaction arrival rate and synchronized period of the data replication, (2) combination of processing time and average CPU occupation rate with and without replica servers under various period of the data replication., and (3) average processing time of transaction at three replica servers under combination of transaction arrival rate and synchronized period of our proposed data replication.