Imagine how the web and the services running on it would look like if sending packets was like sending a letter, that piece of paper you put in an envelop and stick a stamp to it. Priced in a similar model, you would pay for the size, weight and the priority in which you want your packets to be delivered in.
There will be no need to imagine if net neutrality will change and the Internet “fast lanes” come to life.
Stacey Higginbotham from Gigaom wrote a great post about how the airline industry’s fee model for bags, seats, food… is not really payment for better service but actually, degradation of basic service to a level that makes travellers’ experience so miserable that many will prefer to pay extra to avoid it. The limitations imposed on broadband services might take us in the same route as we experienced in air travel.
Impact on innovation
For many of the OTT services we use today (video streaming, internet radio and music and video communication) broadband and their ability to send their media to end users without needing to pay for priority was an important pillar in their capacity to start their business and succeed although they were very small and bootstrapping.
Using asymmetric business models they are not able to charge per usage but rather are typically required to build a significant user base before able to monetize. Being able to send media over the Internet in reasonable QoS without being required for deep pockets is mandatory for enabling new OTTs to innovate.
Similar to the airline industry, allowing companies to pay for higher priority for their service over their competitors will eventually lead to bad service for those not paying.
Why is this important for WebRTC
The whole idea behind WebRTC is that it allows web developers to easily add real-time voice, video and data communication features to their web application and service. The cost of doing this is dramatically reduced with WebRTC because you get a WebRTC implementation in the browser and you are free from worrying about royalties and patents (no need for deep pockets). Given the delay sensitivity of real-time communication, once some traffic sources will be discriminated, real-time communication services will not be able to settle for the basic middle seat, no food, no checked-in bags package on the Internet aircraft but rather will need to pay to play. This in turn will lead to bad user experience and failure of many services. Thus, innovation will suffer.
Real-time traffic is a (very) small portion of the Internet traffic. How could a specialized network service aiming at improving real-time performances of a small portion of the overall traffic could possibly degrade the performances of best-effort traffic ?
ISPs have a strong incentive in providing the best possible performances for Internet services and I don’t see how they could allow themselves to degrade Internet performance: they would just loose customers. The natural market pressure (in markets with clear ISP competition) pushes in that direction and not the other way round.
Regarding the “innovation” argument, I believe ISPs have an interest to see competition emerging especially in front of dominants players like the GAFAs. Where would be the ISP interest in killing innovation ?
Hi Stephane,
The problem is both for semi-real-time and real-time. When one gets higher priority someone else gets lower priority as bandwidth and minimal delay are limited resources.
There is a reason why a content provider like Netflix pays for better service and there is a reason why voice is prioritized in VoLTE networks.
Of course service providers have a strong incentive to provide the best service possible (the airline companies as well). The thing is that if this market becomes a “perfect market” with the aid of SDN there will be those who pay and get better quality.
That is why regulators should keep the right balance between allowing service providers to make money from the bandwidth they provide while not killing the news YouTube or WebRTC service.
Hi Amir,
What Netflix pays is for peering capacity (badnwidth) not prioritization (minimal delay).
Real-time interactive services need minimal delay but are not going by anyway to have a significant bandwidth weight.
So yes, in real-life prioritized (low-delay) VoIP is perfectly compatible with Internet streaming services (bandwidth hungry). AFAIK no-one is complaining about bad Netflix quality because ToIP is being prioritized (which is the common practice for Telephony).
Sure regulators need to keep the right balance for the digital economy to develop to its maximum. To achieve that they need to select relevant fights (fair tax regimes for global Internet players; consider the new monopolies; empower users regarding their own personal data; …) instead of fighting phantoms.
IMHO, in competitive ISP markets (such as in Europe), there is no chance that ISPs could exploit specialized network services to confiscate the revenues of the other players. The market pressure wouldn’t allow for that.
Stephane,
Competitiveness sometimes leads to strong pressure on revenue and therefore creates the need for new revenue sources. That is what happened in the airline industry and what is happening in the ISP market. Question is, what regulations will be created to protect the smaller companies.
While we are here. A nice read about Netflix – http://qz.com/256586/the-inside-story-of-how-netflix-came-to-pay-comcast-for-internet-traffic/
Here is another really interesting view on the events you are quoting: http://blog.cimicorp.com/?p=1684
It’s also good to understand the behaviours of Netflix to push the FCC in some direction: http://cbit.org/blog/2014/09/netflix-secretly-holds-subscribers-hostage-to-gain-favorable-fcc-internet-regulations/
And finally I like to remember how much the fairness arguments are relative: Netflix is really not playing the “Open CDN” game: https://gigaom.com/2014/05/25/will-cable-operators-cdns-and-isps-make-or-break-the-future-of-online-streaming-video/
Finally, AFAIK there are only two ISPs paid by Netflix: Comcast, Verizon and AT&T. In the end it’s really a question of market power and what US operators were able to do is not applicable in Europe which is much more competitive ISP wise. Acutally we have seen the exact opposite move with ISPs fighting to be the first ones to strike deals with Netflix.
Cheers,
Stéphane
On the other extreme, while it’s great to have choices and competition, we currently see a multitude of iterations of similar services because the low barrier to entry, (e.g. opening an S3 or Digital Ocean account) allows anyone to do it. Imagine if The Voice (which I don’t watch) let anyone have a slot on the show?
I’m not for gatekeepers, or budget airline models, but there is another side. I guess the difference is that the Internet is big enough for everyone and in the end, the market decides. Meanwhile, there are 700 WebRTC conference platforms whose web pages all use the same buzzwords. I really wonder how anyone can want to try anything but GoTo Meeting or Hangouts?
Randy,
As you said, the market will decide which services will live and which will fade in the sunset.