Archive for August, 2011
Developments in Europe . . . Benoit Felten, A Slap in the Face of Net Discrimination Lobbyists:
Under the title Are Traffic Charges Needed to Avert a Coming Capex Catastrophy?, economist Robert Kenny builds a systematic refutation of the AT Kearney paper. Kenny disects each of the arguments that forms the AT Kearney reasoning and breaks each one of them down with clinical precision.
The starting point of Kenny’s piece is potentially the most important one: that the need for a change in traffic management is taken as a postulate by AT Kearney and in no way demonstrated. This to me is the most important messages for policy makers and regulators: before meddling with internet traffic management, make sure you understand exactly what is happening, don’t take anyone’s word for it.
From the Introduction to Robert Kenny’s rebuttal:
The net neutrality debate is now gathering steam in Europe, both at the Commission level and in member states. Against this background, four European telcos commissioned a report from AT Kearney [ATK], to support their opposition to net neutrality regulation. This report, A Viable Future Model for the Internet, claims that carriers are facing ballooning capex requirements to fund the growth of internet traffic and that the best way to address this structural problem is via traffic charges to online service providers [OSPs].
If massive capex is required, and this needs to be recovered from OSPs, that would be a significant argument against net neutrality regulation, since it would necessarily end the principal that consumers could access any (legal) site they wished – ISPs would block access to sites that had not paid the charges the ISPs had chosen to impose.
Broadly the logic of ATK’s report as follows:
- Telco investors are already seeing lower returns than investors in other players in the internet value chain
- Telcos face ballooning capex
- This capex is unsustainable
- OSPs are not contributing to the costs of traffic
- In a two-sided market, both sides pay
- Traffic charges are necessary because otherwise OSPs have no incentive to constrain traffic costs
- OSPs can easily afford increased charges
- Increasing retail prices will be challenging
- It is practical to implement traffic charges to OSPs
- Enhanced quality services can be introduced without degrading the basic internet
However I believe both its starting assumptions and its logic are open to significant challenge. This paper reviews the ATK report, from technical, economic and regulatory perspectives, and makes the case that ATK’s conclusion (that the best way forward is traffic charges to OSPs) is not at all well-founded. I consider in turn each of the logical steps above.
Note that the focus of the economic analysis in this paper is primarily on fixed networks, though the qualitative arguments apply equally to both fixed and mobile networks.
(Original at Connected Planet)
What if the car could be used to create a network? What if it could connect to other cars to form a constantly morphing mobile mesh network that helped drivers avoid accidents, identify traffic jams miles before they encounter them and even act as a relay point for Internet access?
By Kevin Fitchard
August 10, 2011
[. . .]
Ford believes the key is Wi-Fi, but not the ordinary access point and receiving device setup. What Ford envisions, [Chief Technology Officer and Vice President of Research Paul Mascarenas] said, is a high-powered, heavily encrypted Wi-Fi that establishes point-to-point connections between cars within a half-mile radius. Those connections could be used to communicate vital information between vehicles, either triggering alerts to the driver or interpreted by the vehicle’s computer. An intelligent car slamming on its brakes could communicate to all of the vehicles behind it that it’s coming to rapid halt, giving the driver that much more warning that he too needs to hit the brakes.
But because these cars are networked—the car in front of yours is connected to the car in front it and so forth—in a distributed mesh, an intelligent vehicle can know if cars miles down the road are slamming on their brakes, alerting the driver to potential traffic jams. Given enough vehicles with the technology, individual cars become nodes in a constantly changing, self-aware network that can not only monitor what’s going on in the immediate vicinity, but across a citywide traffic grid, Mascarenas said.
[. . .]
But Mascarenas said Ford and other automakers can build other applications into the intelligent vehicle network. For instance, not cars but the roads and structures cars use can be embedded with Wi-Fi radios allowing drivers to connect with parking garages, tollbooths or even rest areas through the ad hoc network.
[. . .]
The key, Mascarenas said, is drawing a sharp line between the vehicles as nodes on the network and the vehicles as receivers of information. In order for the system to work, every car acts as a node on the network, occasionally receiving information and services pertinent to the driver but most often acting as a mere relay passing that data down the line of cars until it reaches its destination. When paying a toll, no driver wants to share his credit card data with the 20 cars between his and the tool booth. Ford, however, believes it can put the security and encryption in place that allows such relays to work without compromising individual the privacy of its customers.
The collaborative mesh network could even be used as a mobile broadband alternative to the wide area cellular network. Offload points on the roadside would be used to backhaul traffic to the Internet, but the cars themselves—so long as they all remained a half mile from one another—could pass a Netflix movie stream or a video call down the highway to the vehicle requesting it.