System for distributing an audiovisual content

An elementary module intended to be used in at least one workflow of a system making it possible to distribute audiovisual contents stored or produced live by sources to display terminals, each content being received by a terminal in the form of a succession of consecutive files, referred to as segments, each segment being distributed to a terminal following a transmission of a request by said terminal and being obtained by an application of a workflow to a portion of a content, the elementary module executing a predefined type of processing of the workflow, the elementary module comprising:

• a variable plurality of processing units available for executing the processing of the predefined type, the elementary module regularly obtaining information representing each processing unit added or removed in the plurality, each processing unit not keeping any information concerning a processing once the processing has been performed;
• a module, referred to as a central module, comprising a communication module able to receive requests for first data, each request comprising information representing a workflow to be applied to a portion of a content corresponding to a segment requested in a request by a terminal in order to obtain said first data and information representing configuration parameters to be applied to at least one processing unit of said elementary module in order to generate the first data, each processing unit taking into account the information representing the configuration parameters included in a request for first data for performing a processing corresponding to said request for first data;
• able to transmit a request for second data to a unit preceding the elementary module in the workflow determined from the information representing the workflow, said second data corresponding to said portion processed or not by one or more units preceding the elementary module in the workflow;
• and able to transmit first data when they are available to a unit that sent a request for said first data;
• a scaling module, able to determine a number of processing units to be allocated for implementation of a set of processing operations requested of the elementary module in requests for first data, the determination of said number relying on a scaling model corresponding to the processing of the predefined type supplied by a modelling module (8) included in said system;
• and
• a load balancing module able to choose, for each request received, at least one processing unit from among the processing units allocated by the scaling module for processing second data in order to obtain first data requested and launching the processing of the predefined type by each processing unit chosen, the choice of the processing units relying on a load balancing model supplied by said modelling module;
• each 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 that were 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.