Atos – use case

Context. ATOS provides a platform for live webcasting of sport events such as Masters Tennis. Webcasting of sport events is far more automated and dynamic than in the past and currently end-user feedback and interaction in real time are an integral part of the system.

End-users (i.e. spectators) want to consume the contents in a usable, fast and reliable system that runs on any device or platform (e.g. desktop computer through their browser, mobile devices through a specific app). This makes critical the dependency of the ATOS webcasting system on IT technologies that are continuously evolving, such as mobile and tablets platforms, and background technologies (e.g. Adobe Flash). This dependency has to be incorporated into the webcasting service and its companion player application evolution. Another important feature of webcasting services is to support dynamic adaptation to fulfil end-users expectations. One common way to achieve this adaptation is through Content Delivery Networks (CDN). Although CDNs enable to scale the webcasting service there are two major problems: first, ATOS cannot rely on one CDN (even the most important CDNs cannot guarantee 100% availability); second, worldwide webcasting service has to deal with geographical issues (e.g. CDNs with better infrastructure in different geographical areas, CDNs not allowed to operate in particular countries). These issues raise the complexity of the service adaptation, especially if we consider how critic is adaptation for this kind of systems. For example, in order to reduce the overload of webcasting services generated by the large numbers of end-users, the player application has to include as much logics as possible. This type of architecture (heavy client-side) enables to scale the demand and relies on the CDNs for the worldwide distribution. ATOS needs to provide a system able to adapt to an unknown amount of users, and to different size events, and at reasonable costs.

Of course, the evolution and adaptation of the system is not solely interest of the end-user, from the content producers and distributors perspective, dynamic adaptation of webcasting services is highly demanded, while from the development perspective feedback from end-users is especially relevant in periodic events which represents an important input to the system evolution (e.g. taking into account what was popular or unpopular, or what was more/less effective).

Apart from the evolution and adaptation, the real added value of a webcasting service resides in the personalization of the experience. Different user preferences, needs and constraints need to be considered (e.g. end-user bandwidth, device, location, language, etc.).

Challenges. ATOS would like to improve QoE of the webcasting services offered through their platform, by empowering the platform itself. This means enabling a more dynamic system evolution life-cycle; adding new adaptability capabilities to the platform without damaging its reliability; and providing personalized content and service to the end-users.

Expected benefits and impact. Adopting SUPERSEDE’s methods and tools will enable ATOS to:

  • Improve the webcasting platform evolution. Provided that end-users are an active participant of the software evolution process, ATOS will be able to produce a more competitive set of services for webcasting sport events, and also the final product will be more suitable and appealing for the end-users since it will take into account their suggestions for the webcasting platform.
  • Reach higher-level of adaptability. Considering both the end-user feedback and the monitored context information will bring new possibilities to react in runtime in order to adapt the webcasting system to a variable amount of users, resolve availability problems, and in general increase the reliability of the system.
  • Enable webcasting service dynamic personalization. SUPERSEDE tools will provide fundamental analysis of information that could be used to produce a more advanced webcasting system able to provide a personalized experience for each particular user. Altogether, evolution, adaptation, and personalization, will increase substantially users’ QoE.