Study of algorithms to define the cabling plan of switched Ethernet for real-time applications

Revisiting the notion of Origin-Destination Traffic Matrix of the Hosts that are attached to a Switched Local Area Network

“…Also, let x 1 , x 2 , x 3 , …, x m-1 , x m represent the maximum delay of any data packet through switch 1, the maximum delay of any data packet through switch 2, the maximum delay of any data packet through switch 3,…, the maximum delay of any data packet through switch m-1, the maximum delay of any data packet through switch m respectively, then (11) can be written for any switched local area network as in (12). On the surface, (12) seems to support the notion that there is an origin-destination pairs traffic matrix with respect to end-to-end delay computation for all the hosts that are attached to a switched LAN as enunciated (we think, not correctly) in [1], [2], [6]. We now explain why this notion does not seem to be correct when applied to switched LANs.…”

“…No matter the number of hosts that are attached to the LAN, we can calculate the number of end-to-end delays (number of origin host to destination host delays) with respect to (1) by using (6).…”

A genetic algorithm with switch-device encoding for optimal partition of switched industrial Ethernet networks

“…In the last few years there has been an increasing interest in migrating industrial networks from specific technologies to Ethernet (Skeie et al, 2006;Krommenacker et al, 2001Krommenacker et al, , 2002Lee et al, 2006;Skeie and Johanssen, 2002;Rüping et al, 1999;Georges et al, 2006;Erturk, 2005;Silvestre and Sempere, 2007;Ö zcelik and Ekiz, 2007;Cena et al, 2008;Zhang and Zhang, 2007;Ferrari et al, 2006). Specific technologies currently used, called fieldbuses, such as FIP or Profibus (Georges et al, 2006;Erturk, 2005;Silvestre and Sempere, 2007;Ö zcelik and Ekiz, 2007) are deterministic networks which try to ensure end-to-end delays of messages remain limited.…”

“…GAs (Goldberg, 1989) are computational search algorithms which apply natural-based ideas, such as natural selection, and the survival of the fittest individuals. GAs have been applied to a large variety of network design problems (Kershembaum, 1993;Chu et al, 2000;Salcedo-Sanz and Yao, 2004;Salcedo-Sanz et al, 2006;Abuali et al, 1993;Charddaire et al, 1995), and also they have been found very well suited to tackle the IENPP: (Krommenacker et al, 2001(Krommenacker et al, , 2002 described a first approach to this problem using GAs. These preliminary works adopted a main tree topology, including a redundant line topology.…”

A design process of switched Ethernet architectures according to real-time application constraints

“…As the graph partitioning is an NP-complete problem the Genetic Algorithms are suitable to solve this network segmentation problem. The Genetic Algorithms have already been used for networks performance optimisation (Krommenacker et al ( , 2001). The objective is to collect on the same switches the devices which exchange information.…”