System for distributing an audiovisual content

An elementary module configured to execute a predefined type of processing in a workflow of at least one processing operation implemented by in a system for distributing an audiovisual content stored or produced live by a source to a display terminal, distributing said audiovisual content comprising distributing segments to said display terminal following a terminal request transmitted by said display terminal and resulting from an application of the workflow to a portion of said audiovisual content asked for by the terminal request and identified by an identifier, the elementary module comprising at least one processor configured to:

• receive at least one first data request for first data, said first data request comprising (1) said identifier identifying said portion, (2) information representing the workflow of at least one processing operation to be applied to the portion of the audiovisual content asked for in the terminal request in order to obtain said first data, said information representing the workflow comprising for each processing operation an identifier composed of an address and a port number and (3) a configuration identifier and values of configuration parameters to be applied to said at least one processing operation for generating the first data;
• extract the information representing the configuration parameters contained in said first data request and transmitting a second data request for second data to a source or to a preceding elementary module determined from information representing the workflow, said second data corresponding to said portion of the audiovisual content identified by said identifier after being processed by said source or said preceding elementary module;
• determine a number NU of processors to be allocated for executing said at least one processing operation of the workflow represented in the first request for first data received, said determining using a scaling model corresponding to the processing of the predefined type supplied by a modelling module included in said system, each processor belonging to a plurality of processors available for executing the processing of the predefined type, the elementary module regularly obtaining information representing each processor added to or removed from said plurality, compare the number NU with a number NAA of available processors already allocated and allocate a number of processors equal to NU-NAA in the case where NU>NAA or deallocate a number of processors equal to NAA-NU in the case where NU<NAA;
• choose at least one processor among the processors allocated using a load balancing model supplied by the modelling module and activating each processor chosen so that it applies the predefined type of processing on the second data on reception of the second data, each processor taking into account the information representing the configuration parameters corresponding to the processing that it must execute,
• each scaling and load balancing model being a neural network of the deep learning type, trained by the modelling module during the functioning of said system from metrics obtained by a metrics aggregation module included in said system, said metrics being supplied by each elementary module of said system executing the processing of the predefined type, said metrics representing a functioning of each elementary module of said system executing the processing of the predefined type;
• and,
• transmitting the first data when they are available;
• each processor not keeping any information concerning a processing once the processing has been performed
• wherein the modelling module and metrics aggregation module are software, and wherein the elementary module comprises circuitry to execute instructions.