WoPANets

Use WoPANets

MIFDAOUI Ahlem — 2016-09-23T12:31:42Z

WoPANets embodies the methodology, illustrated in the following Fig.

The main steps of this implemented methodology are :

System Specifications
This first step consists in collecting the maximum of available information concerning the application profiles and the network architecture.
- The application profiles can be defined based on events traces or more commonly on traffic contracts, which define the main characteristics of the exchanged traffic flows, e.g., packet lengths, period or inter-arrival time, deadline, traffic class, jitter, burst size, source, set of destinations, path(s) (...).
- The network architecture is described (if known), e.g., switched networks, wired/wireless links or gateways (if any), and more information on the node behavior and technological characteristics, e.g. service policy, transmission capacity or memory constraints, can be integrated ;

System Modeling
This second step is of utmost importance since it conditions the tightness of the guaranteed quality-of-service metrics. The main idea of this step is :
- First, to model the specified application profiles, and more precisely the maximum exchanged traffic data between applications, based on the concept of maximum arrival curve in Network Calculus framework.
- Then, the system behavior is modeled and particularly the minimum availability of the system guaranteed to the input traffic, i.e., the minimum service curve in NC framework.

Performance Analysis
After building the traffic and system models, the performance analysis of the communication system can be conducted to compute various quality-of-service metrics, such as maximum bounds on end-to-end delay, buffer utilization within crossed nodes, network utilization at the bottleneck or the jitter.

Performance Optimization
Once the performance analysis of the first specified solution is conducted, there are two offered possibilities for the designer, depending on the performance analysis results :
- If the specified solution does not fulfill the system constraints, one may look for an admissible solution guaranteeing these requirements, by defining some variable system's parameters to allow the exploration of alternative solutions ;
- If the specified solution is feasible, one may look for a better solution to enhance the guaranteed system performance, e.g., minimizing the delay or maximizing the network utilization.

Specification Refinement
this last step consists in defining the most accurate specification of the communication network under design, which improves the system's performance while guaranteeing its constraints. This choice depends on the performance optimization step results, and the designer experience to select the most adequate solution among the proposed ones.

Contact

MIFDAOUI Ahlem — 2016-09-23T12:26:38Z

For new projects, training, licensing or any other requests, please contact :

Ahlem Mifdaoui
Associate Professor : Université de Toulouse/ ISAE
Institut Supérieur de l'Aéronautique et de l'Espace (ISAE)
10 avenue Edouard Belin - BP 54032 - 31055 Toulouse cedex 4
Tél. +33 5 61 33 8590 - Fax. +33 5 61 33 91 88
email : ahlem.mifdaoui@isae.fr
Home page : http://personnel.isae.fr/ahlem-mifdaoui/

Partners

MIFDAOUI Ahlem — 2016-09-23T12:26:24Z

Technical Documents

MIFDAOUI Ahlem — 2016-09-23T12:25:24Z

Conference Proceedings

Mifdaoui, Ahlem and Ayed, Hamdi WOPANets : a tool for WOrst case performance analysis of embedded networks. (2010) In : International Workshop on Computer-Aided Modeling Analysis and Design of Communication Links and Networks (CAMAD), 03-04 Dec 2010, Miami, United States.
Ayed, Hamdi and Mifdaoui, Ahlem and Fraboul, Christian Frame Packing Strategy within Gateways for Multi-cluster Avionics Embedded Networks. (2012) In : 17th IEEE International Conference on Emerging Technologies & Factory Automation (ETFA ) 2012, 17-21 Sep 2012, Krakow, Poland.
Ayed, Hamdi and Mifdaoui, Ahlem and Fraboul, Christian Interconnection optimization for multi-cluster avionics networks. (2013) In : 25th Euromicro Conference on Real-Time Systems (ECRTS 2013), 9 July 2013 - 12 July 2013 (Paris, France).
Ayed, Hamdi and Mifdaoui, Ahlem and Fraboul, Christian Hierarchical Traffic Shaping and Frame Packing to Reduce Bandwidth Utilization in the AFDX. (2014) In : The 9th IEEE International Symposium on Industrial Embedded Systems (SIES 2014), 18 June 2014 - 20 June 2014 (Pise, Italy).

Articles

Apvrille, Ludovic and Mifdaoui, Ahlem and Saqui-Sannes, Pierre de Real-time distributed systems dimensioning and validation : The TURTLE method. (2010) Studia Informatica Universalis, vol. 3 (n° 8). pp. 47-69.

Thesis

Mifdaoui, Ahlem. "Spécification et validation d'un réseau de communication de type Ethernet Commuté pour systèmes avioniques militaires de nouvelles générations". PhD, Institut National Polytechnique de Toulouse, 2007
Hamdi AYED : "Optimization and performance analysis of multi-cluster embedded networks", PhD, Institut National Polytechnique de Toulouse, 2014
Dinh khanh DANG : "Performance Analysis of wireless Technologies for new generation Avionics Embedded Systems", PhD, Institut Supérieur de l'Aéronautique et de l'Espace, 2014

Publications related to Avionics Networks

[http://personnel.isae.fr/ahlem-mifdaoui/publications-91.html]

ABOUT

MIFDAOUI Ahlem — 2016-09-23T11:34:27Z

AIM

In the context of critical embedded systems, verification of temporal and functional constraints in the worst-case is an essential property, not only to ensure the proper functioning of the system in its environment, but also and especially to guarantee strict certification requirements, particularly for avionics and space.

The opportunity to make this worst-case performance analysis since the early design phases will allow designers to make important decisions concerning the system's parameters tuning and dimensioning, to avoid wasting time and costs in detailed and erroneous implementation.

The tool WoPANets (Worst-case Performance Analysis of Embedded Networks) provides a first answer to this design challenge. The first applications show that WoPANets reduces development time and costs, while meeting the system requirements since the early design phases.

Key Features

System level-performance evaluation of embedded networks to prove certification requirements
Optimization-based process for design space exploration
Reduction of the design process time and costs
High interoperability thanks to its compatibility with existing design approaches, such as UML and AADL
High level models of the most common real-time networks, i.e., Ethernet, AFDX, and in the near future Spacewire, NoC and Wireless
Easy to use by any designer without any specific knowledge of the used analytical formalism due to an ergonomic GUI
High modularity and scalability to integrate new formalisms and models

Structure

Ergonomic Graphical User Interface to facilitate the use of WoPANets by any network designer, without any particular knowledge of modeling formalisms ;
Analytical methods based on Network Calculus for modeling and conducting worst-case performance analysis, which unlike the formal verification methods allows to overcome the scalability problems ;
Technological libraries of the most relevant real-time networks, i.e., Ethernet, AFDX and in the near future Spacewire, NoC and Wireless ;
Optimization approaches to find an "optimal" setup, which meets the system requirements since the preliminary design stages.