Telecom broadband networks that provide fast lane access to the Internet need to be managed, and even more so than the plain old telephone service circuit-switched networks that provided dial-up. The underlying reason is that so many different services have to be supported, and prioritized to keep the network from congestion, from “real-time applications first, streaming applications second, interactive applications third, and background applications last”, as spelt out by a recent publication by the Information Technology and Innovation Foundation: Managing Broadband Networks: A Policy Makers Guide. It’s a slightly odd publication because it purports to confront net neutrality advocates who are portrayed as in denial. Maybe in the US some are, but it seems on the face of it almost a non-issue. And significantly a completely different issue from carriers (telecom or cable carriers) wanting to charge at both ends of the market: charge the customers for access, and charge the content vendors, such as Netflix, for delivery in return for higher speeds and network prioritization.
So what are the issues? There seem to be three:
i. should broadband be a universal (i.e. public) service, and if so how to set the minimum bitrates available;
ii. is network management transparent and really only used to optimize the utility of the network to serve the maximum number of customers, or is it being used to throttle as a commercial weapon and;
iii. should service providers be allowed to charge at both ends of the market.
Underlying all of these concerns is the question: how competitive is the market? What choice do users really have?
In November, after losing Congress to the Republicans who seem mostly against net neutrality regulations, President Obama reiterated his earlier stance in favour of declaring broadband Internet connectivity a public service. The implication being that a licence to offer broadband Internet connectivity comes with obligations on a commercial enterprise. If the economics don’t work out, and in some locations they may not, then it’s back to some form of universal service fund, or other measures such as opening the local loop. Back to the future as it were. Except that with traditional non-telecom and non-cable companies, such as Internet companies, entering the market the options for more flexible regulation would seem to be greater.
Network management is the thorniest issue to regulate because only the carriers themselves really know their networks. The key issue is how to determine when these companies are using network management as a smoke screen. And in this regard it is worth making the distinction between cyclical patterns of demand, such as the peak loading problems of “busy hour” which need network management, and the secular upward trend in the demand for bandwidth which require investment. Scott Jordan and Arijit Ghosh of the University of California, Irvine have suggested four criteria could be used by regulators to judge whether or not network management practices raise red warning flags:
1. If the management tools are being applied in the transit between the different networks (source network and the carrier network) and in the routers below the transport level this should raise a red flag.
2. If access control involves blocking or termination as opposed to rectifying quality of service degradation this should raise red flag.
3. If the application of management tools is a unilateral decision of the ISP and not a response to the user, for example a content vendor, then this should raise a red flag.
4. The management tool can be applied to (i) an application, (ii) the source/destination, (iii) the service provider, and/or (iv) the payments processor. Tools applied to traffic on the basis of (ii) or only to traffic based on (iii) should raise a red flag.
What is useful about this framework is that it is not deterministic because the red flags are only alerts that can help regulators.
Finally, charging both ends of the market is an ideal solution for commercial carriers, but not for anyone else. Logically it pushes at least part of the burden of financing high-speed broadband network rollout onto the providers of apps and content because if Netflix or anyone else wishes to reach customers in faraway places or even in poor inner city areas, they have to pay a premium to incentivise the ISP to provide the network. It’s not a bargain that will make much commercial sense to the likes of Netflix, but it shifts the burden of responsibility for network build-out to “the market” and that means liability to market failure.
There are only two logical answers to the dilemma. Either find ways to increase competition from new entrants without the legacy of historically high costs, or to declare broadband Internet a public service and finance it accordingly, which then takes us full circle.