Sandvine, a leading provider of intelligent broadband network solutions for fixed and mobile operators, published a whitepaper today entitled "Exposing the Technical and Commercial Factors Underlying Internet Quality of Experience."
In the paper, Sandvine uses data measured from live, representative networks to explain how Internet Quality of Experience is the product of technical and commercial decisions made by the full range of players in the Internet chain, including access network providers, tier-1 and tier-2 IP transit providers, hosting and CDN providers, device manufacturers and Internet standards bodies. As a result of this complexity, over-simplified and popularized "quality" benchmarks that purport to measure access network quality, actually report results that are as influenced (or more) by the decisions of others in the Internet chain, all acting rationally in their own best interests. Decisions of one player can impact end-to-end quality irreversibly and push costs onto the next player.
Some interesting conclusions from the whitepaper include:
- Internet speedometers are broken
- Speedtest and the Netflix Speed Index don't actually measure nor reflect access network speed.
- In the measured networks, YouTube consistently experienced a quality downtime
- A commercial decision by YouTube has resulted in a noticeable quality downtime for the service during lunch and peak evening hours. Other video services don't experience a similar phenomenon.
- Adaptive video, like Hulu and Netflix, slows web, gaming and VoIP traffic
- On busy networks, adaptive video can expand and push aside non-adaptive content such as web browsing, gaming, and VoIP, due to its interaction with TCP congestion control (an Internet standard).
Sandvine has also published on http://www.betterbroadbandblog.com the first in a multi-part blog series of easy-to-digest whitepaper highlights that illustrate some of the complexities in the tussle over Internet quality.