Performance of the SANA-Implementation

In this Page, the Performance of the Implementation is discussed. Especially, the Performance and Scalability of Simulations with lots of Network Nodes or artificial Cells are considered.

The Focus during Implementation was on the Expandability and not on Performance. Thus, each Component is a Java-Object and each Component contains a Status-Method which is called every Time-Step. In this Status-Method, it is possible to perform certain Tasks like e.g. sending Messages, moving to another Node, etc. However, each Status-Method of each Object must be handled every time-step.

Question 1: How many Objects exist in the Simulation?

Question 2: How costly is a call of these Status-Methods?

Summary

The Computational Power scales mostly with the Number of Components (esp. Network Nodes and artificial Cells) in the Simulation and the Number of Packets. This Behaviour is substantiated by the Running-Times of Simulations.

"Performance of the SANA-Implementation" is mentioned on: SANA_AIS