HSC welcomes all external visitors to this site, especially students and members of the academic community. Please use the comments box at the bottom of each page to record any comments or suggestions for improvement.
Network Modeling refers to abstractions of real or imaginary worlds we create to understand the network(wired, wireless and satellite) behavior, play with them ("what if" experiments), make projections, animate or simply have fun.
Computer simulation is the process of implementing models on computers. To do this, we must have appropriate simulation tools.
Network Simulation and Modeling emphasizes on the performance evaluation of computer systems and communication networks. The performance of a system influences its design, procurement and subsequent use. Performance analysis may be carried out at any stage of that life-cycle. In general, the problem is to achieve the highest possible performance given the constraints on the system. Typical constraints might be the number of users, the bandwidth or the cost.
Modelling and simulation are the methods widely and commonly used by performance analysts to represent constraints and optimise network performance.Simulations raise many new questions about appropriate levels of detail in simulation models. Designing simulations to study a protocol inherently involves making choices in which protocol details to implement or use.
The advantage of computer simulation lies in its abilty to model the dynamic network. Unlike the simplistic analytical simulation which is based on mathematical model characterized by set of equations computer simulation which is based
analysis that characterizes a network as a set of equations, the discrete event computer simulation can model the dynamic nature of a network. Discrete event simulation allows to model real events for systems with complex architectures and topologies, test their new ideas and evaluate performances and trade offs.
Simulator can provide abstraction of multiple threads of control and inter-thread communication. Simulators may use finite state machine for process modeling. The processes could represent a function or a protocol stack. Alternatively, the functions and protocols could be implemented by usual program files without any finite state machine. Most of the simulators support both top down and bottom up design methodology.
Simulator typically comes with a set of predefined modules and user-friendly GUI. Some network simulators even provide extensive support for visualization and animation. Thus with most of the simulators the network can be build using drag and drop facilty. Parameters and variables associated with the building blocks of the network can be tweaked to optimize and verify the network performances. Alternatively, simulators may provides the user to build the network from scratch with or without using the inherent modules/libraries.
Figure below shows the Simulation and modeling cycle
Define--->Build--->Simulate--->Analyze&Verify--->Validate
Existing simulators such as OPNET provide detailed protocol implementations
Comments
