NUBOMEDIA: FP7/2007-2013 GA-610576

What's NUBOMEDIA

Multimedia communications
95%
Elastic Clouds
90%
Super-scalability
90%
Computer Vision
60%
Augmented Reality
60%
HTML5 & WebRTC
50%
Multisensory
40%
Social communications
40%
NUBOMEDIA in 30 seconds

The NUBOMEDIA vision

NUBOMEDIA is the first cloud platform specifically designed for hosting interactive multimedia services. NUBOMEDIA exposes to developers PaaS APIs for creating media pipelines: chains of elements providing media capabilities such as encryption, transcoding, augmented reality or video content analysis. These chains allow building arbitrarily complex media processing for applications. As a unique feature, from the point of view of the pipelines, the NUBOMEDIA cloud infrastructure behaves as a single virtual super-computer encompassing all the available resources of the physical network.

The NUBOMEDIA mission

NUBOMEDIA mission is to democratize interactive multimedia communication services by making their creation, deployment and mass-scale exploitation a cheap, rapid and effortless process. To achieve this, we use a strategy composed of two axes. First, NUBOMEDIA exposes its capabilities through a simple to use and intuitive API that can be used by non-expert developers on most popular client platforms such as smartphones and WWW browsers. Second, the NUBOMEDIA infrastructure is released using a flexible and attractive Free Open Source Software license guaranteeing openness and neutrality.
Why to create a multimedia cloud?

The stringent requirements in terms of latency, jitter and packet loss make general-purpose clouds inappropriate for real-time interactive multimedia services, and no specific clouds providing suitable technology solving these issues exist. It is true that there are a several popular person-to-person communication solutions such as Skype, Google Hangouts or Apple FaceTime , which expose PaaS APIs to developers simulating cloud behaviour. However, these services are based on managed network architectures (e.g. P2P) and their infrastructures are private and cannot be customized to special requirements. In general, the creation and maintenance of such interactive multimedia communication infrastructures is unaffordable for most companies, which cannot create customized mass-scale services supporting new features such as augmented reality, computer vision or multisensory information integration.

In this context, we believe that many European companies, and in particular numerous SMEs, could improve their competitiveness and innovation capability if they could create and maintain, without large investments, their specific-purpose interactive multimedia infrastructures in areas such as security, training, e-Health, CRM, ERPs, TV production and broadcasting, logistics, games, social networks, etc. This argumentation allows introducing the main hypothesis of this project: We hypothesize that, in current state-of-the-art, there is a lack of cloud technologies specifically adapted to real-time interactive multimedia. We believe that this lack is an important deterrent to European innovation. We defend that an open solution in this area will provide interesting business opportunities.

Why to create new multimedia APIs on top of that cloud?

Developers, companies and organizations face many different problems to create and exploit interactive media services. The most remarkable of them is complexity. Complexity on the management of the infrastructures, complexity on the development of applications, complexity of integrating novel interactive schemes, complexity for adapting to new socialization schemes, etc. Complexity is at the root of the problem tree and it generates difficulties that slow down innovation such as large prototyping and market times, unaffordable innovation investments, necessity of human resources with specific development and engineering expertise, difficulties to scale successful services and business models, etc. NUBOMEDIA proposition is to minimize the complexity of infrastructures by creating a specific purpose cloud platform bringing all the cloud advantages to the arena of real-time interactive multimedia. NUBOMEDIA will be capable of abstracting and managing all the low level details of service deployment, management, and mass-scale exploitation. In addition, this cloud will host interactive multimedia capabilities that shall be exposed through a simple API, so that the complexity of creating applications on top of the cloud is also minimized.

The NUBOMEDIA concept

NUBOMEDIA is an elastic scalable cloud platform specifically designed for real-time interactive multimedia services, which exposes its capabilities through a simple to use PaaS API. NUBOMEDIA keystones are the following

  • NUBOMEDIA is based on Free Open Source Software (FOSS) and will be released using flexible FOSS licenses. This shall guarantee that the platform is open and can be openly accessed.
  • To defeat fragmentation, NUBOMEDIA shall be designed to simplify the creation of multiplatform live interactive multimedia applications, with special emphasis on WWW and smartphone/tablet platforms.
  • In NUBOMEDIA, the concept of multimedia will be generalized to introduce more than audio and video on the interaction channel. It will provide support for data exchanged from any kind of sensors.
  • NUBOMEDIA will provide seamless integration and access to advanced media capabilities such as Video Content Analysis (VCA - aka Computer Vision), Augmented Reality (AR), media privacy and protection, media monitoring and auditing, media adaptation and media transformation; allowing enriching the information and the interaction capabilities of applications. Immersive and 3D formats will also be supported in the infrastructure.
  • NUBOMEDIA will be able to integrate with third party live interactive multimedia technologies including VoIP systems, telecom infrastructures, CDNs, WWW and TV broadcast systems.
  • The services we will be supporting are, in a wide sense, interactive. This means that users shall be able to interact through the media but also to interact with the media. As a result, we shall support traditional conversational (suitable for maintaining conversations) person-to-person and group communications usage scenarios, but also person-to-machine (e.g. non-linear interactive media browsing), person-to-media (e.g. recognizing a face by touching onto it on a smartphone screen) and machine-to-machine (e.g. starting/stopping recording when a specific event occurs) interactions.
  • Our cloud notion includes capabilities located on the client side. This means that part of the PaaS APIs exposed to developers shall be directly accessible from client devices (e.g. smartphone, WWW, etc.) through proxies. These proxies will constitute a complete SDK abstracting client side complexities associated to media capture, rendering and management.