Read the instructions to get the latest distribution of WETMO from the download page.

Introduction

In the sphere of Differentiated Services performance evaluation, the Telecommunication Networks Group of Politecnico di Torino in cooperation with the NEC Network Laboratories in Germany has developed one additional module that, working with the Network Simulator ns-2.1b8a, allows to simulate web traffic with flows of different lengths exchanged among clients and servers (requests) or conversely (replies). All with network topologies whose complexity is limited only by the memory disposal of the processor the simulator runs on.

Just for its additional functions, this module was called WETMO (WEb Traffic MOdule).

Since the target of WETMO is to simulate the behavior of a DiffServ network with different traffic types or load conditions, the most general network scenario is the one with a central DiffServ subnetwork and various TCP source linked on both sides of DiffServ network.

Furthermore WETMO allows one to collect TCP connections with some common parameters, in aggregates (or clouds). So, managing the traffic coming from clouds in different ways, it is easy to compare the average performance obtained by clouds, or by TCP connections of different lengths in the same cloud. Moreover one can add to web traffic some FTP long lived flows, so as to overload one or more critical routers to observe their behavior with heavy load.

There are many parameters to properly set the simulations, among the others one can tune:

• network complexity
• load of each cloud
• total offered load
• service level agreement
• active queue management (RED algorithm only)
• simulation timing

In the same way, many simulation results are displayed and collected in some files each time an event occurs in the simulation run.

Furthermore in the WETMO package there are three new Traffic Conditioners, planned to be used in the edge routers of DiffServ networks. These new Traffic Conditioners, called Fair Markers (I, II and III), are realized with the target to obtain a scalable device that equalizes the bandwidth resources sharing among the active flows.
Some further details on the Fair Markers usage and characteristics can be found on the Fair Markers manual.

References

• Massimo Alzati, Marco Bottigliengo, Claudio Casetti, Marco Mellia, Fair Marking of Web Flows using Partial State Information, International Teletraffic Congress (ITC-18), Berlin, Germany, September 2003
• Dario Rossi, Claudio Casetti, Marco Mellia, A Simulation Study of Web Traffic over DiffServ Networks, IEEE Globecom 2002, Taipei, TW, November 17-21, 2002
• Massimo Alzati, Marco Bottigliengo, Study and Simulation of Fair Markers in DiffServ Networks, Master Thesis, July 2002
• Dario Rossi, A Simulation Study of Web Traffic over DiffServ Networks, Master Thesis, November 2001