Studies started in year:
 Projects and Studies


Efficient and scalable distribution of on-demand streaming content
For further information please contact:

Adam Kapovits
Programme Manager
Wieblinger Weg 19/4
69123 Heidelberg, Germany

Project Information
What is this Project about?

The background for this project

Providing on-demand streaming content with high quality requires a high bandwidth, even with new coding standards like MPEG-4. It has traditionally been regarded as a unicast service, with separate streams per receiver. This is because each receiver will request the content at different times, and broadcast/multicast cannot be used directly as for traditional TV distribution. With this distribution model, a high penetration will put a high burden on servers and on the infrastructure.

During the last decade, several research projects have been working on finding more scalable approaches. Many different distribution techniques have been proposed that may reduce the total bandwidth demand for the most popular content considerably. These include caching, P2P, and using multicast for on-demand content. Two examples are Skyscraper Broadcasting and Bandwidth Skimming (there are many variations on that), which use multicast in innovative ways to deliver popular content that has several requests for the content during a play-out period. Other distribution methods are also possible, for instance using a shared multicast to download popular content to customer equipment in advance, and storing it for later use. This fits well with a subscription based business model (like a TV series subscription), or when the customer equipment may temporarily store the content even when it is not explicitly requested by the customer.

The introduction of more scalable distribution techniques may change the competitive landscape for ISPs, network operators and content providers. The techniques may reduce the cost associated with distributing popular content dramatically for content provider as fewer servers will be needed. The ISP and network operator may also benefit from this due to the lower demand for bandwidth. Furthermore, the incremental cost of adding receivers is reduced as the number of receivers grow, which may clearly influence the pricing model for such content. It is also important to note that with the most scalable methods, even low-end systems (like standard PC’s) would be able to distribute such content to a large population if (a small) waiting time is allowed before play-out begins at the client side.

IP multicast may also be used in P2P networks to increase scalability, with a potential to provide efficient distribution of high bandwidth on-demand content.

It is also possible to use P2P without multicast. It is of interest to see how the scalability of P2P unicast compares to the other techniques.

The more advanced techniques also have their downsides. First of all the techniques add complexity and require new functionality in client end systems. Some of the schemes require that clients download with a higher receive rate than the play-out rate, and some require that the user waits for a period of time before starting to view the content. Client interaction also gets more complicated (especially “fast forward”).

The study will analyse the different approaches to on-demand content distribution over an IP network. The advantages and disadvantages of each method will be described.

The study will also look at what implications the use of scalable distribution methods for on-demand content may have on the business strategy for ISPs, network and content providers.

What are the main objectives of this Project?

The main objectives of this Study are to:

  • Evaluate different approaches for efficient on-demand content distribution.
  • Evaluate the potential impact for ISPs, network– and content distributors.

What are the key results of this project?

For the private market Video on Demand is expected to be an important service provided by Telecom operators and content providers. TV-distribution will also be an important service, and with a more flexible distribution platform TV-programs may also be distributed in an on-demand fashion. With this in mind it can be argued that most of today’s streaming/broadcast content to private households can potentially be distributed as on-demand content.

The results of this study will help operators making decisions regarding market strategy, and when planning and designing on-demand streaming services.

What relationship exist to other organisations, bodies or initiatives?


Project Results