MediaCloud a Framework for Real-time Media Processing in the Network

, enhancements, and adoption of services will be much more dynamic than today, calling for flexible Software as a Service business and operational models. II. STATE OF THE ART Up to now ever increasing Internet capacity demands could be accommodated by simply enhancing the installed bandwidth. As this concept becomes more and more challenging moving forward [1], there is need for alternative approaches. A standard practice to achieve this goal is to add "higher layer" intelligence to the networks aiming to reduce overall traffic, thus enhancing available capacity by reducing traffic. The first huge success of this concept was the invention and successful introduction of Content Delivery Networks. CDNs basically enabled the massive scale introduction of media services comprising broadcast delivery characteristics in the Internet. At the same time, empowered by mature virtualization technologies, cloud computing infrastructures and services are going to change the IT world and its business models. Basically, cloud computing allows hosting services much more economically, by taking advantage of dynamic use of shared resources. This is a clear advantage for web and enterprise services. But beyond that simple transaction-oriented services there are other comparatively challenging services, which are timesensitive and session-oriented, like multimedia. III. MEDIA IS CHALLENGING Currently we see an emerging trend towards personalized media streams, where media processing is performed somewhere in the Internet, i.e.