NOTE! This site uses cookies and similar technologies.

If you not change browser settings, you agree to it. Learn more

I understand

Learn more about cookies at :

12 January 2016 | Simpler distribution of computing tasks in complex systems

Researchers at List, a CEA Tech institute, have developed a technique for distributing software across machines in complex systems. This novel technique is faster and more efficient than the traditional method.

technique repartition logiciel
© Madgooch / Fotolia

Surprisingly, one of the challenges of complex systems—made up of a large number of software applications that interact locally and simultaneously—is the long, laborious task of distributing the software across the available computing resources. List researchers have solved this problem, with a simple, time-saving approach. In research carried out in partnership with ENS Cachan and CRAN (Centre de Recherche en Automatique de Nancy), List researchers leveraged the institute's Papyrus model-based engineering studio to digitize the many command-control functions of a nuclear power plant.

Once the software bricks were ready, they had to be distributed across the available servers depending on how critical the function they determine is and the capacities of each computer. This "function allocation" step can be carried out by optimization software, which tests all of the possible solutions one by one. However, in systems of systems, this approach is too complex and, therefore, costly in terms of both human resources and time. To solve this problem, the researchers used List's Diversity algorithm validation platform to find a solution that would respond to all of the specified requirements, but without seeking optimization at all cost, "A concession that considerably reduced computation time without compromising on security or reliability."

The approach has been validated for use in the nuclear industry and could now be used to optimize complex systems in general, and, specifically, in advanced manufacturing to distribute the production of a heterogeneous set of products across several factories, for example.