The Experiment

Experiment is a network scenario designed by the users to simulate a desired network behavior. Experiments in eLab are separated into 2 parts.

The first part is designing and creating the experiment itself. The user can use the experiment layout to design a network topology, install applications and protocols on each node and configure the parameters on the links. When the user finish to build the topology he can schedule different events for each installed application.

Figure 1.1 Creating an Experiment

typical_experiement_create.jpg

The second part is to analyze the result with different views. eLab provides different views to help you study what actually happened in the experiment. The different views allows the user to analyze the data in different ways, for example one of the views allow the user to observe packets moving from node to node another view allows the user to check the throughput with a nice chart.

Figure 1.2 Experiment result as packet animation

 

typical_experiement_result.jpg

 

 

Experiment Architecture


The design architecture of eLab is using network simulation as the engine. Network Simulation (NS) projects started back in 1989 and was supported by DARPA, Xerox, PARC, UCB and USC/ISI and its widely supported by researchers all over the world.

The online interfaces provides by eLab are all interacting with ns engine via session management. This design provides a reliable validation to experiment result designed by users.

Figure 1.3 eLab Backend is using network simulation

  elab_workflow.jpg

With this design each user can submit their experiment to the engine and have it stored for future inspect.

Introduction to Nodes >>