Musicalization of a Complex Computational Ecosystem

De
Aller à la navigation Aller à la recherche

Project Name

MUSCOE (Musicalisation of a Complex Computational Ecosystem)

Coordinators

Participants

Starting date

February 2014

Project Description

The idea is to monitor a complex computing ecosystem in a musical way rather than via a visual way.

Complex computational ecosystems may comprise several hundred computers that working in common towards a common task (trying to find a complex mathematical model behind some observed data, for instance). A large enough problem may require very long runs, lasting for several days.

Using a visual interface to monitor the run will require operatos to watch a screen for many hours, which can be very tedius, knowing that:

  1. it is impossible to do anything else while you visually monitor a run possibly because
  2. abnormal behaviour may be transient, meaning that if a problem is punctual, one has to watch the screen exactly when the problem happens.

Rather than relying on vision, this project proposes to use hearing as the primary means for detecting abnormal behaviour among the computing ecosystem. If something weird is detected through audition, the operator may then open a window to visually have a look at what is actually happening. The audio monitoring then serves as a warning that something strange is happening.

The original idea behind this project is to use Ligeti harmonics for several reasons:

  • Ligeti was using mathematical and physical concepts to create his music, so using a complex system to create Ligeti harmonics may not be far from Ligeti's original thoughts
  • Ligeti is the music that was used by Stanley Kubrik in his marvelous movie 2001 a Space Odissey, featuring a spaceship automatically operated by a massively parallel super-computer called HAL. The analogy between HAL and a computer grid or a cloud of computers is very strong, so having a complex computational ecosystem compose eerie music à la Ligeti would be strongly and strangely reminiscent of HAL.

However, the first prototype is currently distorting known music, rather than composing new music using mathematical laws. First results are quite promising as can be heard here:

The project has been presented during a CNRS Mastodons Mesures-HD meeting on April 29th 2014 as well as during a JET evolutionary workshop on June 20th 2014. A paper describing this project will be submitted to EVOPar 2014.

Acknowledgements

This project is funded by the French ANR Emergence EASEA-CLOUD project.

Image:Astronotus_ocellatus.jpg|Astronotus ocellatus (Oscar)