FLAME: FACILITY FOR LARGE-SCALE ADAPTIVE MEDIA EXPERIMENTATION

FLAME instead takes a holistic approach with the platform reaching from the top-level experiment level over the traditional compute & storage cloud infrastructure down to the routing and transport network level, providing experimental control over aspects at all levels. Due to this approach, the outcome of a FLAME experiment aims at providing a deeper understanding of how to manage OTT media applications and deliver insight into platform operations for every experiment that is executed, seeding commercialisation of the results with the gained insights. To achieve this, FLAME offers the following capabilities:

  • Declarative specification language for infrastructure-level resources (compute, storage and network) consider- ing structural and behavioural requirements of the experiment
  • Foundation media services for experimentation such as: elastic video management, efficient video transcoding and processing, adaptive video streaming and delivery, fast upload and ingestion, replication and synchronisation of services, service monitoring, general media repository, compression-focused media encoder; including use of resource elasticity (computing and network) within an experiment;
  • Resource reservation with QoS to ensure required performance characteristics (e.g. service response time, net- work latency, etc.) necessary for media service in real world situations at city scale
  • Surrogate service endpoint provisioning and management exploiting distributed multi-cloud virtualized re- sources of different scales for minimizing service response time and latency
  • Flexible routing as a service solutions based on ICN and SDN techniques for highly dynamic network topologies aligned with content and demand characteristics.
  • Fine-grain physical, virtual and application monitoring, and other performance and utilisation metrics for the overall platform and applications (e.g. orchestrator, topology, etc);
  • Media service QoS/SLA supervisor to explore relation with end-users QoE 1 and enable the necessary control over experiments
  • Online analysis of experiment data for media and content optimisation, providing the necessary insights from the experiment data;
  • Management of the entire experiment lifecycle, including a API as well as an experiment and monitoring toolbox to define and manage experiments;
Share on FacebookTweet about this on TwitterShare on LinkedInEmail this to someonePrint this page