Skip to main contentIBM Zurich Research Laboratory
 
IBM
Home IBM Research IBM Switzerland      
 Select a country
 IBM Research Home
Zurich Research Lab
Computer Science
· Completed projects

DABWeb: Datacasting via digital radio

Imagine for a moment that you have a lightweight portable device that you can use not only for browsing the Web but also for reading your local and national newspapers, for browsing through your favorite e-commerce catalogs (and buying that cool new gadget you just read about on that cool IBM Research Web site). Go a step further and imagine that the information on the device is always up-to-date and—best of all—that you do not have to be connected to Internet! Nor do you have to mortgage your house to pay for costly mobile phone connections. You think we are dreaming? Well, maybe we are (for, after all, dreaming is part of the research process)—but we are not delusional. Let us explain.

Most of the data that we obtain via the Internet today is delivered in so-called pull mode: The user (using a Web browser) has to go to the Web site she is interested in and load the page(s) into her computer; that is, she has to pull the data to her system. Each Web access is a one-to-one connection to some content provider (in fact, it often is even a four-to-one connection as a browser typically opens up four concurrent connections for a single page). Once the data has been downloaded to her computer, the user has to reload the page to get an up-to-date version.

Looking at this way of disseminating data via the Web we can identify the following issues:

  • The user has to be fully connected to the Internet.
  • Downloaded content can only be refreshed by the user, not by the content provider.
  • The content provider has to command bandwidth that is proportional to the content size times the number of concurrent users.

If we consider the situation from a mobile computing perspective, we have the additional constraint that mobile bandwidth is limited and priced at a premium. Thus, yes, your scepticism is well-founded: Distributing content that requires large bandwidth does not appear to be something you would want to try with the current technology.

But wait! How about reversing the distribution mechanism? Instead of using a pull-model-based approach, why not use a push mechanism? And instead of using precious mobile bandwidth we radio broadcast the data? Remember the radio? That technology from the previous century to distribute music, talk shows, plays, sports, and what-have-you over the air to billions of receivers worldwide? Well, that is the technology we are using—but with a twist: All over the world radio is going digital.

The most advanced system, both with regard to deployment and to the features offered, is probably the European Eureka-147 digital audio broadcasting (DAB) system that is in operation commercially all over Europe as well as in the Far East (Singapore, Taiwan), Australia, and Canada. In the US there is the terrestrial in-band-on-channel DAB (IBOC DAB) system, promoted by iBiquity, and the satellite digital audio radio system (SDARS) by XMRadio and Sirius Radio. All these systems have in common that they broadcast digitally, and some of them—Eureka-147 DAB and IBOC DAB—explicitly offer data channels.

Being located in Europe and with Eureka-147 DAB being the most advanced system, we based our work on Eureka-147 DAB. In the DABWeb project we have set up a complete inhouse DAB testbed consisting of a DAB transmission system and various receivers. We have developed a Java-based DAB stack that allows us to utilize the two datacasting vehicles offered by Eureka-147 DAB: Program-associated data (PAD) and data channels. The PAD feature allows us to piggyback data onto a broadcast audio stream. Applications for this include sending the cover image of the currently running soundtrack but also for scrolling text. The data channel provides a high-bandwidth (up to 384 kb/s per channel) delivery mechanism to the receivers.

Our DAB framework is modular and can easily be extended and adapted to new appliations (allowing us to experiment) and new receivers (and transmitters). Part of the framework is a content-based subscription mechanism that utilizes the IBM JavaCard. Using a smart card we implemented a conditional access scheme that does not require a backchannel to some authorizing server. We have successfully implemented a Broadcast Web Site application where we transmitted complete Web sites to DAB receivers, certain Web sites required the user to have a suitable subscription applet on her JavaCard.

Ongoing and future research will explore DAB with a backchannel: Quite a number of interesting and exciting applications become possible with even a small bandwidth backchannel such as the SMS feature of the GSM networks.

WAP phone

     
  About IBM  |  Privacy  |  Terms of use  |  Contact