In order to test fault-tolerance of the system, we experimented in a network, which includes 800 DHT nodes. We simulated node’s failure. We used a simple node failure model, which considers the crashing of node or the network failure. In the node failure model, all the failure nodes are simulated simultaneously. After simulated node’s failure, we immediately sent successively lookup requests to the available DHT nodes, and we counted the number of failed lookup requests and total requests and calculated the lookup failure rate along with the node failure rate. The value of lookup timeout in this experiment is 10 seconds. Figure 11 shows the result of lookup failure rate and node failure rate. The lookup failure rate increases fast if the node failure rate is more than 40%. If the node failure rate is less than 40%, the lookup failure rate is small (