Periodic Reporting for period 2 - FLAME (Facility for Large-scale Adaptive Media Experimentation)
Période du rapport: 2018-01-01 au 2020-09-30
FLAME is developing a ground-breaking platform for large-scale experimentation of FMI services fully integrated with broadcasting supporting high mobility scenarios and use of software-defined and cloudified network infrastructures including mobile edge computing for optimised communication, storage and service provisioning. The objectives are to R&D
1. A ground-breaking platform for large-scale experimentation of FMI services fully integrated with broadcasting supporting high mobility scenarios and use of software-defined and cloudified network infrastructures including mobile edge computing for optimised communication, storage and service provisioning
2. A methodology for real-life ecosystem experiments investigating acceptance (QoE), performance (QoS) and viability that addresses constructs of personalisation, interactivity, mobility, and localisation
3. A cross-layer control plane for investigating OTT media service workflows with controllable and replicable conditions
4. Mechanisms to establish FLAME Replicators offering EaaS for investigations within FMI ecosystems implemented at different European cities
5. A pioneering large-scale FMI experiments investigating novel PIML content production and delivery workflows within a range of socio-economically important vertical areas
6. EaaS sustainability and business models that combine public-private investment in localised infrastructures with an exploitable software platform
1. A ground-breaking platform for large-scale experimentation of FMI services fully integrated with broadcasting supporting high mobility scenarios and use of software-defined and cloudified network infrastructures including mobile edge computing for optimised communication, storage and service provisioning
2. A methodology for real-life ecosystem experiments investigating acceptance (QoE), performance (QoS) and viability that addresses constructs of personalisation, interactivity, mobility, and localisation
3. A cross-layer control plane for investigating OTT media service workflows with controllable and replicable conditions
4. Mechanisms to establish FLAME Replicators offering EaaS for investigations within FMI ecosystems implemented at different European cities
5. A pioneering large-scale FMI experiments investigating novel PIML content production and delivery workflows within a range of socio-economically important vertical areas
6. EaaS sustainability and business models that combine public-private investment in localised infrastructures with an exploitable software platform
FLAME has laid down the technology foundations necessary to deliver and launch the alpha EaaS based on the FLAME platform. Initial definition of requirements, experimentation methodology, architecture, and systems integration processes necessary to implement and test the platform through integration, staging and production have been completed. D3.1 provides an initial vision for the FLAME Future Media Internet (FMI) through scenarios and use cases. D3.3 describes the first architecture of the FLAME platform including detailed specifications of platform capabilities and components driven including Service Function Routing, Service Function Endpoint Management and Control, Orchestration and Cross Layer Management and Control (CLMC). D3.5 describes the roadmap for development, integration and production deployment of the platform including a devops environment with automated testing.
FLAME has defined an initial methodology for real-life ecosystem experiments investigating acceptance (QoE), performance (QoS) and viability that addresses constructs of personalisation, interactivity, mobility, and localisation (PIML) as described in D3.2. The methodology targets acceptance and viability of Future Media Internet (FMI) systems in real-life settings considering the expected demand patterns of PIML workflows.
FLAME has made significant progress towards design and implementation of a cross-layer control plane for investigating media service workflows with controllable and replicable conditions. The FLAME architecture elaborates the concept of CLMC and the interplay between configuration, measurement and actions taken by the platform to orchestrate and control routing and endpoints. The alpha release includes Service Function Routing (SFR) and CLMC features, along with foundation media services for metadata management and streaming of video content. These features offer a minimal viable platform to support the validation experiments and their deployment within the urban infrastructures of BIO and i2CAT.
FLAME is working to extend OASIS TOSCA for spatially and temporally provisioned services along with the higher-level KPI constructs needed to allow media service providers to define performance targets. A first version of the experimentation toolkit has been implemented providing a monitoring and measurement model for time-based data integration and analysis of cross-layer information. A first version of the FLIPS platform supporting service function routing has been delivered.
FLAME has documented a set of best practices, technical descriptions and guidelines that aim to ease replication of a FLAME-capable infrastructure in D5.1. The descriptions and guidelines are based on the current status of a production environments (Bristol) as well as integration plans, preliminary testing and deployment results for the Barcelona infrastructure. The replication process considers the business, technical, deployment, and tooling aspects.
FLAME has made progress towards supporting pioneering large scale FMI experiments. The focus has been on elaborating and designing the validation experiments from high impact verticals of broadcast, gaming and transmedia as a way of building the evidence for technical performance, user acceptance and viability. FLAME has implemented a marketing and communication strategy, plans to engage FMI ecosystem stakeholders and has created an exciting FLAME website. FLAME has made progress in preparing for the open call which plans to launch Apr-18. FLAME has engaged with the FIRE community: FLAME was presented as part of the data management session in 2nd FIRE engineering conference and will be a key note speaker at the 3rd FIRE Engineering Conference in Paris. FLAME has engaged with key market players and has a strategy towards media, 5G and cities through stakeholder workshops. FLAME has completed market analysis to position FLAME within the FMI ecosystem. The market analysis was used to explore the FLAME value proposition and its exploitation and sustainability models.
FLAME has defined an initial methodology for real-life ecosystem experiments investigating acceptance (QoE), performance (QoS) and viability that addresses constructs of personalisation, interactivity, mobility, and localisation (PIML) as described in D3.2. The methodology targets acceptance and viability of Future Media Internet (FMI) systems in real-life settings considering the expected demand patterns of PIML workflows.
FLAME has made significant progress towards design and implementation of a cross-layer control plane for investigating media service workflows with controllable and replicable conditions. The FLAME architecture elaborates the concept of CLMC and the interplay between configuration, measurement and actions taken by the platform to orchestrate and control routing and endpoints. The alpha release includes Service Function Routing (SFR) and CLMC features, along with foundation media services for metadata management and streaming of video content. These features offer a minimal viable platform to support the validation experiments and their deployment within the urban infrastructures of BIO and i2CAT.
FLAME is working to extend OASIS TOSCA for spatially and temporally provisioned services along with the higher-level KPI constructs needed to allow media service providers to define performance targets. A first version of the experimentation toolkit has been implemented providing a monitoring and measurement model for time-based data integration and analysis of cross-layer information. A first version of the FLIPS platform supporting service function routing has been delivered.
FLAME has documented a set of best practices, technical descriptions and guidelines that aim to ease replication of a FLAME-capable infrastructure in D5.1. The descriptions and guidelines are based on the current status of a production environments (Bristol) as well as integration plans, preliminary testing and deployment results for the Barcelona infrastructure. The replication process considers the business, technical, deployment, and tooling aspects.
FLAME has made progress towards supporting pioneering large scale FMI experiments. The focus has been on elaborating and designing the validation experiments from high impact verticals of broadcast, gaming and transmedia as a way of building the evidence for technical performance, user acceptance and viability. FLAME has implemented a marketing and communication strategy, plans to engage FMI ecosystem stakeholders and has created an exciting FLAME website. FLAME has made progress in preparing for the open call which plans to launch Apr-18. FLAME has engaged with the FIRE community: FLAME was presented as part of the data management session in 2nd FIRE engineering conference and will be a key note speaker at the 3rd FIRE Engineering Conference in Paris. FLAME has engaged with key market players and has a strategy towards media, 5G and cities through stakeholder workshops. FLAME has completed market analysis to position FLAME within the FMI ecosystem. The market analysis was used to explore the FLAME value proposition and its exploitation and sustainability models.
Progress beyond SOTA during the period
• FLAME offers an experiment toolbox through features of the Orchestrator (TOSCA specification) and the CLMC (monitoring). Through TOSCA templates and infrastructure abstractions systems under test can be controlled and monitored independently of physical location.
• FLAME’s architectural infrastructure abstraction based on common standards and specifications, along with integration and testing across different Infrastructure Products ensures technical replication at different locations
• FLAME’s KPI and dimension information model within the CLMC allows for measurement and analysis of cross layer function and performance. The focus on KPIs and dimensional measurements in the experiment design provides the use cases and scopes testing of measure procedures for different characteristics of the system under test
• FLAME extends the TOSCA specification to include additional constraints that can be used to orchestrate media services at different geo-graphic locations. Features for TOSCA++ templates, management, control, monitoring and analytics features support the full lifecycle of experimentation. This covers features of the Orchestrator, CLMC, SF Endpoint Management and SF Routing.
• SF endpoint management and SF routing features support reconfiguration of SF instances in response to surrogate policy constraints. The Orchestrator's TOSCA++ templates allow for re-specification of constraints. The CLMC’s configuration interface supports notification of failure events associated with media service and SF lifecycles.
Impacts:
• Enlargement of experimental facilities coverage: alpha components implemented and integration started. FLAME launch scheduled for 1 March along with plans to join the FED4FIRE federation
• Total number of experiments running on the FIRE facilities: 4 validation experiments in broadcast and gaming vertical sectors
• FLAME offers an experiment toolbox through features of the Orchestrator (TOSCA specification) and the CLMC (monitoring). Through TOSCA templates and infrastructure abstractions systems under test can be controlled and monitored independently of physical location.
• FLAME’s architectural infrastructure abstraction based on common standards and specifications, along with integration and testing across different Infrastructure Products ensures technical replication at different locations
• FLAME’s KPI and dimension information model within the CLMC allows for measurement and analysis of cross layer function and performance. The focus on KPIs and dimensional measurements in the experiment design provides the use cases and scopes testing of measure procedures for different characteristics of the system under test
• FLAME extends the TOSCA specification to include additional constraints that can be used to orchestrate media services at different geo-graphic locations. Features for TOSCA++ templates, management, control, monitoring and analytics features support the full lifecycle of experimentation. This covers features of the Orchestrator, CLMC, SF Endpoint Management and SF Routing.
• SF endpoint management and SF routing features support reconfiguration of SF instances in response to surrogate policy constraints. The Orchestrator's TOSCA++ templates allow for re-specification of constraints. The CLMC’s configuration interface supports notification of failure events associated with media service and SF lifecycles.
Impacts:
• Enlargement of experimental facilities coverage: alpha components implemented and integration started. FLAME launch scheduled for 1 March along with plans to join the FED4FIRE federation
• Total number of experiments running on the FIRE facilities: 4 validation experiments in broadcast and gaming vertical sectors