Live VoD exists today as IPTV. It works about as well as regular VoD does, provided it is not a “Spectacle” event, like a Super Bowl, or a World Cup.
The network’s fragility and waste are most in evidence when the spectacle demands critical paths and massive redundant bandwidth. Spectacle Live video will unavoidably trend towards UHD resolutions.
If the internet, using Brute Force Unicast [BFU] wishes to keep up, it has material issues of scale that it will need to address. A few stats regarding the TV audience of major events. (Giuliano & Gallo, 2019)
- Average NFL game: 15M, (A staggering 600 Tb/s at 4k UHD)
- Super Bowl: 100M
- 2015 Cricket World Cup: 200-300M
- 2018 FIFA World Cup Finals: 500M
The trend for Live Streaming appears suitably robust:
- Viewers spend 8X longer with live video than on-demand (5.1 minutes for on-demand vs. 42.8 minutes for live video content). (50Wheel, 2018)
- Live video is outpacing the growth of other types of online video; a 113% increase in ad growth yearly. (Yahoo News, 2016)
The network’s fragility and waste are most in evidence when the spectacle demands massive redundant bandwidth. Spectacle Live video will unavoidably trend towards UHD resolutions.
The foremost Content Delivery Network [CDN] is Akamai. Akamai currently sources roughly 20% of the world’s total Internet traffic. Akamai’s peak delivery via Brute Force Unicast [BFU] is on the order of 70 Tbps. This represents only 1.8M concurrent viewers at 4k UHD (Holland, 2019). Ominously, that’s a tiny fraction of the numbers shown above.
Not surprisingly, Live Super Bowl “broadcast” over the internet has, to-date, been a case study in disappointment.
The paper “Did the Superbowl Kill the Internet” (Morales & Bergstrom, 2017) illustrated the failures consequential of the dynamic nature of audience volume changes, which are apparently rather sizable.
Forty percent of these failures lasted 50 seconds or longer. 60% were for 20 seconds or longer, 66% were routing path failures, the remaining 33% were not.
A route gets congested, forcing a search for another less congested path and withdrawal of some of the original routing. Regrettably the new path can have path commonalities with the first congested path, and the new route cannot handle the dynamic traffic increase. Part of the original route recovers due the earlier withdrawal, and traffic is redirected to the earlier pathways, causing the whole process to repeat. This aberrant behavior is termed “BGP Flapping”.
This is where Multicast methods offer the economy, and performance that unicast cannot offer. These methods also provide a unique economy for emerging Live video markets. Connected Fitness applications like Peloton come to mind.
A Live distribution modality should be integrated into any Hybrid SVoD initiative.
Given the anticipated demand for live content, it would be irresponsible not to.
Hybrid VoD 