There is no one definition of Net Neutrality. It is a concept rather than a technical or legal definition. It is not even clear in many eyes whether it applies to all networks (no) or just the Public Internet (yes) – and what happens where they both share infrastructure such as broadband access (err… good question).
Certain things are definitely “not neutral” under any definition, such as deliberate degradation or outright blocking of legal Internet content or applications, on otherwise uncongested networks. Conversely, other aspects such as capped data plans, the use of CDNs or protections against denial-of-service attacks are viewed as completely acceptable, by all but the most hardcore “activists”.
Much of the current furore is around the prospect of paid-priority “fast lanes”, otherwise known euphemistically as “specialised services”, and whether they force other non-partnered traffic onto unusable “slow lanes”. This also tends to draw distinctions between fixed and mobile data broadband, because the former already have dedicated QoS-managed pathways for carrier VoIP and IPTV, while mobile broadband generally does not, except for a few tough-to-implement VoLTE instances.
Given this blog’s focus, the main emphasis here is on mobile data models rather than fixed broadband.
The vagaries of mobile networks such as interference, mobility, devices & coverage, and the fact that most traffic to smartphones goes via 3rd-party WiFi anyway, tend to make discussion of general 4G paid-QoS mostly irrelevant. You can also add in the lack of any obvious willingness of market participants to pay for “unprovable but slightly better than average QoS, honest” and the ever-better capability of most software to deal with occasional network variability and glitches.
Some market participants try to steer discussion towards totally irrelevant topics, such as paid peering arrangements, or “discrimination” by Apple or Google in their app-stores or access to OS APIs – usually to try to create strawman arguments about their own preferred stance on real Net Neutrality. While both might be examples of “unfairness” in an abstract sense, they are completely decoupled from broadband access to the Internet or other services.
Such “red herrings” are simply used to deflect the Neutrality discussion. But a more interesting set of developments can be considered “grey areas”. Many mobile operators are trying to nibble around the edges of Net Neutrality – partly to see how far they can push regulators’ boundaries, but also in an effort to differentiate their broadband propositions. They are testing the waters with approaches such as:
- User-defined prioritisation, where the customer rather than telco defines their own Internet connection’s “non-neutrality”, or chooses ISP-managed offers like “gamer broadband”, or various child-protection content-filtering options.
- Various “worse-efforts” concepts, where “unimportant” traffic can be delayed until, or sent pre-emptively during, network quiet periods.
- Zero-rating of certain content types, i.e. exempting specific data from users’ quotas. (I posted a detailed post here a few weeks ago). In this scenario, nobody pays per-MB for the data, although there may be revenue-share or other partnership models involved.
- Sponsored-data / “sender-pays” models, such as that proposed by AT&T, where (in theory) 3rd-party application or content firms pay to offset users’ traffic consumption. I wrote a fairly scathing post here
- Application-based charging, for example a flat-rate per day for using Facebook or Whatsapp.
- User-based prioritisation, where certain customers (eg business users) receive better speeds or QoS than other users. While this does not explicitly differentiate between applications or content, in most instances businesses don’t watch Netflix, and consumers don’t use SAP.
- Differential treatment or pricing of data that goes via proxy-based platforms (eg Opera Mini, BlackBerry) rather than direct to the Internet
- M2M services managed differently to their consumer-centric data plaforms
- Dedicated MVNO-type deals, such as those used by Amazon with its early Kindles, which look similar to “sponsored data” models.
It can be seen that most of these “grey areas” involve differential pricing, rather than more “invasive” network-based QoS and prioritisation. Such approaches are easier both technically and in discussions with regulators. They also allow the operators’ marketing teams to work (fairly) swiftly to construct plans and deals, without getting mired in the cumbersome technical and organisational realities of network operations.
Two of the most high-profile approaches to “Grey Neutrality” are the zero-rating and sponsored-data approaches. AT&T’s January announcement of its Sponsored Data API programme has been a catalyst, and also in the US recently, T-Mobile has launched its “Music Freedom” plan as well as zero-rating speed-test data.
Orange zero-rates traffic to its personal cloud service for consumers, Telia zero-rates Spotify and DT zero-rates its German mobile TV service. Of course, many other internal operator services such as MMS or closed-garden portals have long been available without incurring extra data usage costs, while BlackBerry plans have often been subscription-based rather than per-MB.
Numerous emerging-markets operators have free offers to access Facebook/Twitter/Wikipedia/Whatsapp/Viber and numerous other services, including some giving free access to educational materials – although often this is in countries which don’t have Neutrality legislation anyway.
To be fair, the original vision for mobile data services was that they would be “on net” and billed in various modes, rather than volume-based Internet access. To a large extent, “raw” Internet access was only offered on mobile after most of the other 3G service models failed to gain any traction.
It is therefore unsurprising to see operators look to go back to the model of bundled or separately-billed services – although the main question is how this fits with the app-download paradigm, which normally implies a need for “vanilla” Internet access rather than hooks into a telco-resident server.
While the lawyers and lobbyists continue to argue over “fast lanes” in a black-and-white fashion, Disruptive Analysis believes that these tend to miss the point anyway, at least in mobile. There are only a few realistic use-cases for such QoS-optimised propositions, and they are unlikely to ever contribute more than a couple of percentage points to total mobile broadband revenues.
The action is going to be in the “grey areas”. Disruptive Analysis new report predicts over 1.5bn users of zero-rating by 2019, for example. In the sponsored data realm, the most promising category is in advertising, where it seems relatively non-controversial (and consumer-friendly) for unsolicited promotional content to be paid for by a third party, rather than consuming the user’s quota.
There may be a few other “grey areas” emerging as well – expect interesting variants to emerge as SDN/NFV takes hold over the next 5 years – perhaps with CDNs or caching to be pushed down to the radio network. This will enable many services to be delivered “not directly from the Internet”, although whether the APIs are available to mainstream mobile app developers is questionable. New forms of MVNO may emerge, as well as capabilities delivered via private cellular networks, WiFi or via telcos’ API platforms.
Regulators and others will find these challenging to deal with, so need to ensure that rules are framed flexibly and clearly.
In any case – they key thing for Net Neutrality isn’t the fireworks at the FCC or in Brussels. It’s the smaller instances of edging closer to the rules’ limits, especially around charging. Conversely, we should expect a lot more push-back and unintended consequences, where operators try to unilaterally block or degrade mainstream web services. We already see Netflix “outing” Verizon’s perceived/alleged games – that type of naming & shaming (and perhaps lawsuits) will be much more common, as well encryption and obfuscation.