Clock synchronization method for network simulation environment of discrete event

Abstract

The invention provides a clock synchronization method for network simulation environment of discrete event. The method comprises the steps of: 1) determining a virtual clock tick count value corresponding to occurrence time of a discrete event when the discrete event occurs, and sending the tick count value as a target count value to a kernel-state program; controlling operation of all timers of real software accessing to a network simulation environment of discrete event by using the virtual clock tick count value; 2) summating the virtual clock tick count value gradually by using 1 as step length to reach the target count value by the kernel-state program; for a real software accessing to a network simulation environment of discrete event, determining whether the timer of the real software overtimes or not each time when the virtual clock tick count value is added by 1. The method of the invention assists in avoiding distortion caused by time inconsistence between simulation environment and real environment. In the invention, the real software can be applied in network simulation environment of discrete event directly without modification.

Description

technical field [0001] The invention relates to the technical field of network simulation and simulation, in particular, the invention relates to a clock synchronization method of a discrete event network simulation environment. Background technique [0002] With the rapid expansion of the network today, network researchers must constantly think about new network protocols and algorithms, and do forward-looking basic research for network development; on the other hand, they must also study how to utilize and integrate existing network resources to make the network reach For maximum performance, both require validation and analysis of new network solutions. The cost of traditional experimental methods is very high, it is difficult to reconfigure or share resources, and it is not flexible enough to use. It is difficult to make the scale of the test bed large, and it cannot realize the integration of various communication flows and topologies in the network. to great limitatio...

AI Technical Summary

Problems solved by technology

Compared with other simulation methods, the discrete event network simulator has obvious advantages in speed and efficiency, and has strong scalability. However, due to the use of discrete events for time-driven, the simulation time is inconsistent with the actual time. In some cases The test results in the simulated environment will deviate from the results in the real environment

For example, when a real software requests a server in the network simulator to establish a TCP connection, if the clock of the discrete event simulator advances very slowly (when there are many events to be processed), the server in the network simulator will The TCP connection of the real software failed due to timeout before the TCP connection was answered

However, when the real software judges that the connection fails, in fact, the clock of the network simulator may only advance a small part, and it has not reached the moment when the real software times out.

Obviously, in this case, the test results in the simulated environment are distorted

Images