With regard to business models, real data has been used from existing network in France and extrapolated to Europe. The order of magnitude of the savings brought by techniques explored by CONVINcE is some hundreds of millions Euros per year in Europe (90 m€ for IPTV and 660 m€ for OTT, respectively 13% and 14% savings at service level). Using a cloud architecture showed slightly lower figures (80 m€ for IPTV and 620 m€ for OTT), which is compensated by the flexibility brought by a cloud architecture.
Video encoders are the main contributors to headend energy consumption. Applying different techniques studied by the project (e.g. algorithms optimization, use of GPUs or edge-cloud transcoders) can reduce energy consumption by a factor of two. It was also demonstrated that using last-generation encoding technology (HEVC) increases the consumption of encoders by 50%, but end-to-end consumption is reduced significantly, especially in the network area, as the bitrate is also divided by 2.
Several solutions were studied to reduce the consumption in networks. A gain of 10%-15% was achieved by using an all-optical core network. Using DOCSIS3.1 appliances for cable network nodes and amplifiers provides up to 45% improvement. Energy-efficient routing algorithm for SDN-based access networks can save up to 10% and load balancing algorithm in SDN-based data center networks up to 40%. Virtualized multi-tier architecture brings up to 93% savings in video surveillance networks. Further potential savings were discovered in CDNs and WiFi access networks.
CONVINcE identified also potential energy savings in terminals. Video playback power consumption can be reduced by 83% with optimized execution environment in terminals. Virtualized offloading (saving up to 32%), fountain coding (22%) and radio resource management (15%) can also reduce terminal power consumption. The highest savings in TV displays (45%) is provided by including new optical film structure, LED efficacy improvements and auto-dimming algorithms. Finally, encrypted communication and network interface selection can give extra reductions.