Actions sur le document


The main objective of the AtNet team is to develop network algorithms and protocols which fit as driver software into network equipments. To produce performance evaluation results, these algorithms and protocols are generally implemented into a network simulator (e.g. NS-2, OPNET, etc). They are discrete event network simulators. They support popular network protocols, offering simulation results for wired and wireless networks alike. NS is popular in networking research given its open source model and online documentation.

Websites have been built to provide an online access to some of our network softwares:

On Internet DNSSEC

Click for details

  • KROd is a program that performs automatic DNSSEC key-rollover and automatic conversion from DNS to DNSSEC.
  • A patch of BIND which modifies the behavior of the DS field. Generalized DS allows to have build a DNSSEC chain of trust over a succession of secure and insecure domains (a domain that has insecure parents).
  • libsresolv is a library built with the BIND toolkit. It comes as a patch over the BIND 9.3 sources. It contains a DNSSEC resolver and validator. The goal is to show anything that can be proved from a DNSSEC answer. The validator proves positive and negative answers (it can prove that a domain doesn't exist), it can also prove that some domain are empty non-terminal ones. libsresolv performs bottom-up validation, it is signature oriented.


On Explicit Multicasting

  • Into the NS simulator, we have included the Xcast protocol according to the Explicit Multicast (Xcast) concepts. Click for details
  • Simple Explicit Multicast (SEM) uses an efficient method to construct multicast trees and deliver multicast packets. SEM is original because it adopts the source-specific channel address allocation, reduces forwarding states in non branching node routers and implements data distribution using unicast trees. Click for details
  • Generalized Explicit Multicast (GXcast) is a generalized version of the Xcast protocol. It permits  Xcast packet fragmentation and support an increasing number of members in a multicast group. Click for details


On Multicasting in MPLS Networks

  • The MPLS Multicast Tree (MMT and it's extension MMT2)  is a new approach to construct multicast trees in MPLS networks. This approach utilizes MPLS LSPs between multicast tree branching node routers in order to reduce forwarding states and enhance scalability. In our approach only routers that are acting as multicast tree branching node for a group need to keep forwarding state for that group. All other non-branching node routers simply forward data packets over traffic engineered unicast routes using MPLS LSPs. Click for details


On Optical Networks

  • In our study OMNeT++ is used to design and simulate multi-band optical networks. OMNeT++ is an object-oriented modular discrete event network simulation framework. It has a generic architecture, so it can be (and has been) used in various problem domains: modeling of wired and wireless communication networks, protocol modeling, modeling of queueing networks, modeling of multiprocessors and other distributed hardware systems validating of hardware architectures. In general, modeling and simulation of any system where the discrete event approach is suitable, and can be conveniently mapped into entities communicating by exchanging messages.
  • We developed a simulator for node and link protection using p-cycles for dynamic multicast traffic in optical DWDM networks. This simulator is implemented in MATLAB.


On Network Monitoring

  • This simulator is written in C++ under Linux. it uses the ILP solver CPLEX for solving integer linear programs, and the topology generator BRITE for generating random test topologies. It ensures the following features: (i) Given an input network topology, it computes an optimal set of monitor locations and an optimal set of detection paths that can detect all potential link-level anomalies, while minimizing the inherent costs jointly. (ii) Given an input network topology, it computes an optimal set of monitor locations and an optimal set of localization paths that can pinpoint unambiguously the localization of all potential link-level anomalies, while minimizing the inherent costs jointly. (iii) Given an input network topology, it assesses the cost and the speed of continuous anomaly localization (detection and localization procedures are ran simultaneously), and the cost and the speed of reactive anomaly localization (the localization procedure is run only upon detecting an anomaly). On the light of this comparative assessment of the two localization approaches, it suggests a localization configuration (localization approach and monitoring frequency) that offers a good balance between cost and speed for the input topology. Note that the simulator computes optimal solutions, when the exact solutions (ILP based solutions) are used. However,  exact solutions are not scalable. Thus, heuristic solutions are used for large topologies.


On Resource Allocation in Wireless Networks

  • The object-oriented programming capabilities of the Matlab language enable us to develop our discrete event simulator for network selection in heterogeneous environments:. The goal is to elaborate an optimized simulation environment where session arrivals, network selection algorithms, traffic generation, and session departures are implemented. Our simulator is used to evaluate the performance of the different network selection methods, and to compare them to our proposed solution.
  • The OPNET simulation platform has been used in order to design and evaluate the performances of our proposals relating to new opportunistic schedulers. They allow maximizing global system throughput while ensuring fairness without any trade-off. In these works, we have had implemented realistic channel model and traffic sources. Click for details