Summary: ‘Net Neutrality’ has gathered increasing momentum as a market issue, with AT&T, Verizon, major European telcos and Google and others all making their points in advance of the Ofcom, EC, and FCC consultation processes. This is Telco 2.0’s input, analysis and recommendations. (September 2010, Foundation 2.0,, Executive Briefing Service, Future of the Networks Stream).
NB A PDF copy of this 17 page document can be downloaded in full here. We'll also be discussing this at the Telco 2.0 Executive Brainstorms. Email firstname.lastname@example.org or call +44 (0) 207 247 5003 to find out more.
In this paper, Telco 2.0 recommends that the appropriate general response to concerns over ‘Net Neutrality’ is to make it easier for customers to understand what they should expect, and what they actually get, from their broadband service, rather than impose strict technical rules or regulation about how ISPs should manage their networks.
In this article we describe in detail why, and provide recommendations for how.
‘Net Neutrality’ is an issue manufactured and amplified by lobbyists on the behalf of competing commercial interests. Much of the debate on the issue has become somewhat distracting and artificial as the ‘noise’ of self-interested opinion has become much louder than the ‘signal’ of potential resolutions.
The libertarian ideal that the title implies is a clever piece of PR manipulation of ideas of freedom of access of information, and freedom from interference. For the most part, this is far from the reality of the motives of the players engaged in the debate.
Additionally, the ‘public’ net neutrality debate is being driven by tech-savvy early adopters whose views and ‘use cases’ are not statistically representative of the overall internet population.
This collection of factors has created a strange landscape of idealist and specialised viewpoints congregating around the industry lobbyists’ various positions.
However, behind the scenes, the big commercial players are becoming increasingly tense, and we have recently experienced a marked reluctance from senior telco executives to comment on the issue in public.
Our position is that, beyond the hyperbole, the fair and proper management of contention between Internet Applications and ‘Specialised Services’ is important in the interests of consumers and the potential creation of new business models.
Rapidly increasing use of the Internet and Specialised Services, particularly bandwidth hungry applications like online video, is causing (or, at least, will in theory cause) increasing contention in parts of the network.
The currently expressed primary concerns of net neutrality activists are that some consumers will receive a service whose delivery has been covertly manipulated by an external party, in this case their ISP. Similarly, some application and service providers fear that their services are or will consequently be discriminated against by telcos.
Some telcos think that certain other large and bandwidth-hungry applications are receiving a ‘free ride’ on their networks, and their corporate owners consequently receiving the benefits of expensive network investments without contribution. As a consequence, ISPs argue that they should be entitled to unilaterally constrain certain types of applications unless application providers pay for the additional bandwidth.
One of the areas of obfuscation in the ‘Net Neutrality’ debate is the confusion between two sets of issues in the debate: ‘moral and legal’ and ‘commercial’.
Moral & legal issues include matters such as ‘freedom of expression’ and the right to unfettered internet access, the treatment of pirated content, and censorship of extreme religious or pornographic materials. We regard these as subjects for the law where the service is consumed / produced etc., but that have in some places become entangled in the ‘Net Neutrality’ debate and which should not be its focus.
The commercial issue is whether operators should be regulated in how they prioritise traffic from one commercial application over another without the user’s knowledge.
Contention can arise at different points between the service or application and the user, for example:
As a service may originate from and be delivered to anywhere globally, the first kind of contention can only be truly be managed if there is either a) an Internet-wide standard for prioritising different types of traffic, or b) a specific overlay network for that service which bypasses the internet to a certain ‘outer’ point in the network closer to the consumer such as a local exchange. This latter class of service delivery may be accompanied by a connection between the exchange and the end-user that is not over the internet – and this is the case in most IPTV services.
To alleviate issues of contention, various ‘Traffic Management’ strategies are available to operators, as shown in the following diagram, with increasingly controversial types of intervention to the right.
Figure 1 - Ofcom's Traffic Management Continuum
Operators already do apply traffic management techniques, an example of which was given by 3UK’s Director of Network Strategy at the recent Broadband Stakeholder Group (BSG) event in London, who explained that at peak times in the busiest cells, 3 limits SS7 signalling and P2P traffic. He explained that they selected these categories because they are essentially ‘background’ applications that have little impact on the consumer’s experience, and it was important to keep down latency so that more interactive applications like Web browsing functioned well. A major ‘use case’ for 3UK was identifying which cells needed investment.
In 3UK’s case, there was perhaps surprisingly more signalling traffic than there was P2P. Though this is a mobile peculiarity, it illustrates that assumptions about problems in managing traffic management can often be wrong, and it is important that decisions should be taken on the basis of data rather than prejudice.
While there are vociferous campaigners and powerful commercial interests at stake, it is fair to say that the streets are not often full of angry consumers waving banners reading ‘Hands off my YouTube’ and knocking on the doors of telcos’ HQs. While a quick and entirely non-representative survey of Telco 2.0’s non-technical relatives-of-choice revealed complete ignorance and lack of further interest in the subject, this does not necessarily mean that there is not, or could not be, a problem, and it is possible that consumers could unwittingly suffer. On balance though, Telco 2.0 has not yet seen significant evidence of a market failure. We also believe that the mechanisms of the market are the best means of managing potential conflict.
We broadly agree with Alex Blowers of OFCOM, who said that ‘80% of the net neutrality debate is in the definition’ at the recent BSG conference.
First, the term ‘Net Neutrality’ does not actually distinguish which services it refers to – does ‘Net’ mean ‘The Internet’, ‘The Network’, or something else? To most it is taken to mean ‘The Internet’, so what is ‘The Internet’? Despite the initial sense that the answer to this question seems completely obvious, a short conversation within or outside the industry will reveal an enormous range of definitions. The IT Director will give you a different answer from your non-technical relatives and friends.
These ambiguities have the straightforward consequence that the term ‘Net Neutrality’ can be used to mean whatever the user wants, and its use is therefore generally a guarantee for mindless circular arguments and confusion . In other words: perfect conditions for lobbyists with partial views.
For most people, ‘the internet’ is “everything I can get or do when my computer or phone is connected online”. A consumer with such a view probably has a broadband line and an internet service and is among those, in theory at least, most in need of protection from unscrupulous policy management that might favour one form of online traffic over another without their knowledge or control. It is their understanding and expectation of what they have bought against the reality of what they get that we see as the key in this matter.
In this paper, we discuss two classes of services that can be delivered via a broadband access line.
1. Access to ‘The Internet’ (note capitalisation), which means being able to see and interact with the full range of websites, applications and services that are legitimate and publicly available. We set out some guiding principles below on a tighter definition of what services described as ‘The Internet’ should deliver.
2. ‘Specialised Services’ are other services that use a broadband line, that often connect to a device other than a PC (e.g. IPTV via set-top boxes, smart meters, RIM’s Blackberry Exchange Server (BES)) or a service that may be connected to a PC but via a VPN, such as corporate video conferencing, Cloud or Enterprise VOIP solutions.
While ‘Specialised Services’ are not by our definition pure Internet services, they can also have an effect in certain circumstances on the provision of ‘The Internet’ to an end-user where they share parts of the connection that are in contention. Additionally, there can be contention between services on ‘The Internet’ from multiple users or applications connected via a common router.
Additionally, fixed and mobile communications present different contexts for the services, with different potential mechanisms for control and management. Mobile services have the particular difference that, other than signalling, there is no connection between device and the network when data services are not being used.
One possible mechanism to improve consumer understanding and standards of marketing services is to introduce a framework for defining more tightly services sold as “Internet Access”.
In our view, services sold as ‘The Internet’ should:
Where a customer has paid extra for a Specialised Service, e.g. IPTV, it is reasonable to give that service priority to pre-agreed limits while in use.
The point of defining such an experience would be to give consumers a reference point, or perhaps a ‘Kitemark’, to assure them of the nature of the service they are buying. In instances where the service sold is less than that defined, the service would need to be identified, e.g. a ‘Limited Internet Access Service’.
To understand the limitations and possible advantages of ‘traffic management’, and put this into context, it is worth briefly reviewing some of the other ways in which customer and service experiences vary.
Different Services Work in Different Ways
Many Internet Services use already use complex mechanisms to optimise their delivery to the end-user. For example:
Equally, most ISPs are able to ‘tune’ their data services to better match the characteristics of their own network. Although these assets are only available to the services that pay for, own, or create them, none of these techniques actively slows any other service. Indeed, and in theory, by creating or using new non-congested routes, they free capacity for other services so the whole network benefits.
Consumer Experiences are Different too
Today’s consumer experience of ISP services varies widely on local factors. Two neighbours (who happen to be on different nodes) could, in theory get a very different user experience from the same ISP depending on factors such as:
These factors will, in many cases, massively outweigh performance variation experienced from possible ‘traffic management’ by ISPs.
Internet Protocols try to be ‘Fair’
The Internet runs using a set of data traffic rules or protocols which determine how different pieces of data reach their destinations. These protocols, e.g. TCP/IP, OSPF, BGP, are designed to ensure that traffic from different sources is transmitted with equal priority and efficiency.
Further Technical Fixes Are Possible
Network congestion is not an issue that appeared overnight with the FCC’s 2005 Comcast decision. In fact, the Internet engineering community has been grappling with it with some success since the near-disaster in the late 1980s that led to the introduction of congestion control mechanisms in TCP.
Much more recently, the popular BitTorrent file sharing protocol, frequently criticised for getting around TCP’s congestion control, has been adapted to provide application-level congestion control. The P4P protocol, created at MIT and tested by Verizon and Telefonica, provides means for P2P systems and service provider networks to cooperate better. However, it remains essentially unused.
A further consideration is that it is necessary to be realistic about what can be expected – we have heard the benefits from traffic-shaping cited as an extension of around 10% in the upgrade cycle in the best-case scenario.
It’s Complex, not Neutral
It is therefore simply not the case that all Internet services progress from point of origin somewhere in the cloud of cyberspace to the end-users via a random and polite system. There are assets that are not equally shared, significant local variations, and there are complex rules and standards.
‘The Internet’ is a highly complex structure with many competing mechanisms of delivery, and this is one of its great strengths – the multiplicity of routes and mechanisms creates a resilient and continually evolving and improving system. But it is not ‘neutral’, although many of its core functions (such as congestion control) are explicitly designed to be fair.
In principle, Telco 2.0 endorses developments that support new business models, but also believes that the rights of end-users should be appropriately protected. They have, after all, already paid for the service, and having done so should have the right to access the services they believe they have paid for within the bounds of legality.
In terms of how to achieve this balance, it’s very difficult to measure and police service levels, and we believe that simply mandating traffic management solutions alone is impractical.
Moreover, we think that creating a fair and efficient market is a better mechanism than any form of regulation on the methods that operators use to prioritise services.
There are three basic ways of creating and fulfilling expectations fairly, and empowering end-customers to make better decisions on which service they choose.
‘Net Neutrality’ or any form of management of contention is not an issue for corporate customers, most of whom have the ability to configure their IP services at their will. For example, a financial services trader is likely to prioritise Bloomberg and trading services above all other services. This is not a new concept, as telcos have been offering managed data services (priority etc) to enterprise customers for years over their data connections and private IP infrastructure.
Some more advanced consumer users can also prioritise their own services. Some can alter the traffic management rules in their routers as described above. However, these customers are certainly in the minority of Innovators and Early Adopters. Innovation in user-experience design could change this to a degree, especially if customers have a reason to engage rather than being asked to do their service provider’s bottom line a favour.
The issue of unmanaged contention is therefore likely to affect the mass market, but is only likely to arise in certain circumstances. To illustrate this we have selected a number of specific scenarios or use cases in which we will show how we believe the principles we advocate should be applied. But first, what are our principles?
There are broadly three regulatory market approaches.
Telco 2.0 advocates a hybrid approach that promotes market transparency and liquidity to empower customers in their choices of ISP and services, including:
The criteria of the assessment would include the actual performance of the operator against claimed performance (e.g. speed, latency), and whether they adhere to the ‘Code of Best Practice’.
How might it work?
The communication of this assessment could be as simple as a ‘traffic light’ style indicator, where a full Internet service meeting best practice and consistently achieving say 90% of claimed performance would be ‘Green’, while services meeting lower standards / adherence or failing to report adequately would be signalled ‘Amber’ or ‘Red’. The principles used by the operator should also be published, though utilising this step on its own would run the risk of the "Licence Agreement" problem for software - which is that no-one reads them.
We’ll be working on refining our guidelines and thoughts on how an indicator or other system might work by working through some specific ‘Use Cases’ outlined below. In the meantime, we recommend the suggestions made by long-time Telco 2.0 Associate Dean Bubley in his Disruptive Analysis’s ‘Draft Code of Conduct for Policy Management and Net Neutrality’.
It is our view that as long as Telcos are forced to be open, regulator (and consumer bodies) can question or, ultimately, regulate for/against behaviours that could be beneficial/damaging.
We believe that the roles of the regulator(s) should be to:
We draw a parallel with what the UK regulator, Ofcom, used to do for telephony:
Additionally, we see the following possible theoretical service layers within an Internet Service that could be used to create new business models:
One possibility that we will be exploring is whether it could be possible to create an ‘On-demand Enhanced Service’. For example, to deliver a better video streaming experience the video provider can pays for their traffic to take priority over other services with the express consent of the customer. This may be achieved by adding a message to the Enhanced Service e.g. ‘Click here to use our Enhanced Video Service where we’ll pay to get your video to you quicker. This may cause degradation to the service to other applications currently active on your broadband line while you are using the Enhanced Video Service’.
We have long thought that there is scope for innovation in service design and pricing – for example, rather than offering a (supposed) continuous 8Mbps throughput (which most UK operators can’t actually support and have no intention of supporting), why not offer a lower average rate and the option to “burst” up to high speed when required? ISPs actually sell each other bandwidth on similar terms, so there is no reason why this should be impossible.
We’ve identified a number of specific scenarios which we will be researching and developing ‘Use Cases’ to illustrate how these principles would apply. Each of these cases is intended to illustrate different aspects of how a service should be sold to, and managed by / for the customer to ensure that expectations are set and met and that consumers are protected appropriately.
Fixed ‘Use Cases’
There is reasonably common ground between most analysts and commentators on the need for more transparency in Internet Access service definition and performance and management standards, though there is little clarity yet in the ways in which this clarity might be achieved.
The area which is most contentious is the notion of ‘non-discrimination’ – that is of allowing ISPs to prioritise one form or source of traffic over another. AT&T are firmly in favour of ‘paid prioritisation’ whereas Google/Verizon are not, and says ‘wireline broadband providers would not be able to discriminate against or prioritize lawful Internet content, applications or services in a way that causes harm to users or competition’.
Interestingly, in the Norwegian Government’s Net Neutrality Guidelines issued in 2009, provision is made that allows operators to manage traffic in certain circumstances in order to protect the network and other services.
Free Press are a US activist movement who champion ‘Net Neutrality’. While we have accord with their desire for freedom of speech, and understand the imperative to create a more level-playing field for media in the US, our position is not aligned in terms of enshrining total neutrality globally by regulation.
In terms of the regulators’ positions, the UK’s Ofcom is tentatively against ‘ex-ante’ regulation, whereas the FCC seems to favour non-discrimination as a principle. The FCC is also asking whether mobile and fixed are different - we say they are, although as the example of 3UK shows, the differences may not be the ones you expect. Ofcom is also already looking at how it might make switching easier for customers.
We also note that US-based commentators generally see less competition on fixed internet services than in Europe, and less mobile broadband options for customers. Our position is that local competitive conditions are a relevant consideration in these matters, albeit that the starting point should be to regulate the market as described before considering a stronger stance on intervention in circumstances of low local competition.
‘Net Neutrality’ is largely a clever but distracting lobbyists’ ploy that has gathered enormous momentum on the hype circuit. The debate does create a possible opportunity to market and measure broadband services better, and that’s no bad thing for customers. There may also be opportunities to create new business models, but there’s still work to be done to assess if these are material.
1. “Internet Access” should be more tightly defined to mean a service that:
2. Where a customer has paid extra for a ‘Specialised Service’, e.g. IPTV, it is reasonable to give that service priority to agreed limits while in use. Services not meeting these criteria should be named, e.g. “Limited Internet Access”.
3. ISPs should be:
4. The roles of the regulator are to: