In February 2019, Cisco published a whitepaper setting out the company’s research into global internet traffic trends. Based on the data collected, Cisco predicted that IP video traffic will constitute 82 percent of all IP traffic (both business and consumer) by 2022. In fact, the volume of statistics that show just how important online video is (and will continue to be) to the average internet user is mind-boggling and it is evident that it will be an increasing burden on current data infrastructures. To add to the problem, existing solutions are severely antiquated, both in terms of hardware and software. Consequently, inefficiencies from processing, compressing, transmitting, and decompressing multi-media data result in significant energy demands. If we are to take climate change seriously, then addressing the contribution of data centers globally to the problems associated with global warming is a worthwhile pursuit. The effects are highlighted in the following observation:

The music video for “Despacito” set an Internet record in April 2018 when it became the first video to hit five billion views on YouTube. In the process, “Despacito” reached a less celebrated milestone: it burned as much energy as 40,000 U.S. homes use in a year.

And that is just the environmental impact for a 5-minute clip alone! According to a recent study by Huawei Technologies Ltd., data centers currently consume about 2% of electricity worldwide but this number is expected to rise anywhere between 8% and 21% by 2030. This is more than the energy consumption of many nations and puts the carbon footprint of data centres globally on par with emissions generated by the global aviation industry. And whilst video streaming is only a component of all data center usage, it is clear that our penchant for watching video content online will soon take the position as the predominant form of data transfer globally. As a corollary, one day it may become the leading cause of carbon emissions on the planet.

Naturally, many companies are racing to find solutions to the problem ranging from faster hardware to running data centers on clean or renewable energy sources. However, these solutions are topical at best and do little to address the core of the problem itself, the compression and decompression mechanics (referred to in the industry as “codecs”) used by streaming video content providers are themselves inefficient. Because the codecs employed by the majority of the industry are based on code-heavy software, their complexity alone requires greater computational power, which in turn increases energy consumption. David Ronca, the director of video encoding at Facebook (formerly at Netflix), makes the following observations about recent advances in codec technology:

“New codec technology is impressive. But each new generation has come with a 6-10x increase in computational complexity […] We must explore ways to meet the explosive growth of video while bending the compute and power requirements down. Future codec research must emphasize compression efficiency AND computational efficiency. Here’s a thought exercise: If the next video encoder maintained the efficiency of VVC/AV1 but reduced computational complexity by 50%, would we consider that a successful new codec?”

This quote highlights a trilemma facing the next generation of codec solutions: is it possible for a codec to compress high-quality images without significant loss of quality in a manner that is not computationally intensive? The reduction of coding complexity is a worthy pursuit. Consider, for instance, that for a one-hour episode of your favorite TV show recorded in the highest possible quality (assuming here 4K), it can take up to 60 hours to run the required compression calculations where the aim is to preserve image quality! It is hardly surprising that the Despacito video, mentioned previously, could be so environmentally damaging. 

Fortunately, a European start-up called POLAR HPC Solutions Limited is hopeful that its new codec (based on an innovative breakthrough in codec mathematics) will result in lower bandwidth use, reduced data storage needs, and greater energy efficiency without sacrificing high video quality. Furthermore, the solution, which involves accelerating the codec computations on proprietary FPGA-based devices, will lower data center operating costs. Interestingly, the benefits also apply to the consumer end. Not only can data centers pass on their savings in terms of lower fees but a smaller file size and lower computational complexity mean that viewing online videos will take less time and therefore require less power. Couple this benefit with the fact that video content is increasingly viewed on mobile devices and the aforementioned advantages together contribute to fewer battery resources consumed per video. Mobile devices can, therefore ‘last’ longer between charges, which not only reduces the individual’s energy bill but also, in the aggregate, contributes to a reduction in energy consumption by society as a whole.

Part 2 will explore the shortcomings of current codec technology and demonstrate how POLAR’s new codec can achieve results others cannot, simply because it is based on a different mathematical foundation. To find out more about how POLAR’s solution for the online video streaming industry can be of benefit to your data centre, please visit: http://www.polarhpc.com