The R&D division is devoted to studying and developing new technologies related to content delivery services through which end-users can improve their overall experience. During last 40 years computer engineering and telecommunications have been an enormous driving force for society and economy all over the world. During last twenty years we have witnessed the rapid transformation of the Internet from a text-based information platform to a multimedia communication platform. Following Moore’s law we have seen a rapid increase of circuit integration, memory capacity and processing speed at more and more affordable prices making the mass production of computers, routers, mobile phones, PDAs and other electronic devices a reality. Likewise, the vertiginous progress in infrastructure technologies based on optical fiber have brought to end users Internet access speeds which were unthinkable only a couple of years ago. Japan has one of the best telecommunication infrastructures in the world with more than 25,000,000 households connected to broadband Internet access services. The advanced infrastructure in Japan makes this country a very interesting place to test new platforms and services like, for instance, those related to live streaming and on-demand video delivery over ad-hoc P2P networks. Bearing all that in mind, the R&D division is currently working on the following areas:

P2P networking

P2P networks have emerged as one of the most radical shifts from the traditional client-server approach for Internet applications. Content delivery based on P2P is one of the most promising solutions to reduce the burden on the servers and help distributing contents in a reliable and timely way. The potential of P2P networking and its topological organization via Distributed Hash Tables (DHTs) is enormous and there is still a lot of space for enhancement, in particular for video-on-demand and live streaming. Copyright-protected contents can be also safely delivered over P2P networks if appropriate DRM technology is utilized.

Network Coding:

relay nodes in conventional networks are just “dumb forwarders”, i.e. they just relay the data towards the next hop previously decided by the routing algorithm. Network Coding allows relay nodes to create linear combinations of data units received from other nodes in the network before sending them off. The idea is simple but extremely powerful. Network Coding can not only increase the throughput in a wired or a wireless network but also has proven to be extremely effective for data dissemination in multi-hop networks, multicast and broadcast in wireless networks, content distribution in P2P networks and even effective for network security.

Novel Transport Protocols for Efficient Audio/Video Transmission:

the audio standard AAC and video codec standard H.264/AVC are making the transmission of good-quality A/V contents over scarce bandwidth channels a reality. The video codec H.264/SVC extends its H.264/AVC predecessor by adding temporal, spatial and resolution scalability through the use of a base layer and multiple enhancement layers. The transport of such parallel layers of encoded data using the conventional TCP or UDP can be very inefficient as we need to protect in a differentiated way each of the coded layers. In this regard, novel transport techniques such as Stream Control Transmission Protocol (SCTP) provide a more flexible and efficient way to transmit such high-quality A/V contents. Along with the surge of video on demand as well as video live streaming, this is surely a topic that will become increasingly important in next coming years.

Social Networking and Data Mining:

the appearance of several applications based on people interconnection and interaction is raising the importance of social networks and their relation with customer profiles and user data bases. The study of such social-network interaction with efficient data management, dissemination and distribution has an enormous potential to bring new insights into the design of novel service platforms that customize their contents according to each user’s behavior.