Guest Post: Charging for Bandwidth - the world is not flat
The broadband incentive problem is causing ISPs to crack all over the world. In this guest post Fergus O'Reilly, CTO of Highdeal, discusses the changes we need to make in our IT systems to deliver new business models.
Broadband operators across Cable, ADSL, Fiber and Wireless networks are beginning to re-examine how they charge for bandwidth.
The flat-rate, one-size-fits-all broadband tariff offer has been a big hit with consumers since it is easy to understand, easy to compare across ISPs and does not require understanding exactly what "bandwidth" is and which services actually consume it. Broadband has seen strong adoption over the past years in part because of that simple pricing model.
But this model is now under challenge:The Pain Points
- Bandwidth usage has risen uncontrollably. Throughout the 2000s actual usage of broadband bandwidth soared. Initially this growth was driven by "unofficial" peer-to-peer downloads among individuals, but now a lot of the growth is from music and video services like Apple iTunes, Google YouTube, BBC iPlayer, and others.
- Operators have discovered that a small proportion of their subscribers consume a very large proportion of their bandwidth. Time Warner Cable, for example, let slip in 2008 that 5% of their consumers use over 50% of their bandwidth.
- When operators sell "unlimited" and "all-you-can-eat" packages they may promise say 20Gpbs to all subscribers but their networks are not sized properly to handle the traffic if suddenly all subscribers each try to use that 20Gbps in parallel.
- The infrastructure was built under the premise of "build it and they will come" and come they did. It would be hypocritical to turn around now and complain that content providers are building too many cool services or that dematerialization of audio and video content is a nuisance.
- To continue to provide quality service, operators have to continually invest in their core, backhaul and peering networks to handle the increased traffic. This investment is very costly, meaning that service provider margins are being severely compromised.
Solution options tried
Get Content Providers to Pay - operators have raised the idea of passing on the costs to the Internet content and application companies.
This idea got rapidly shot down because of fears about Network Neutrality. Content providers are happy to pay their share for connecting their servers to the Internet, but they do not want to have to pay for preferential carriage - they do not want to be put in a position where ISPs discriminate against them unless they pay up.
De-prioritization & Throttling: Silently de-prioritize some high-bandwidth traffic like peer to peer exchanges.
Network Neutrality again. The US FCC rapped Comcast over the knuckles for doing this without properly informing their subscribers.
Cap & Overage schemes: some ISPs, particularly some of the North American Cable providers such as TimeWarner, and Rogers, have started rolling out new tariffs based on cap and overage charging for high bandwidth consumers.
This can be a hard sell since it requires consumers to carefully track their usage. The majority of subscribers have absolutely no idea of how much a Gb of bandwidth is, nor should they.
What needs to be solved
Subscribers have become used to having always-on functionality at predictable pricing, so moving away from the flat-rate model is going to be a hard sell for consumers. Service Providers will need to introduce new models to manage bandwidth and charge accordingly. They need to get more revenue from subscribers who use the most bandwidth, without penalizing or confusing the remaining subscribers.
Some answers to the pain
One way to achieve this bandwidth management is by introducing some form of tiered-pricing based on usage & bandwidth boosts. Service providers will probably strive to introduce more sophisticated schemes, such as noticing when you are trying to upload the family vacation home movie to a video sharing site and offering an instant 10x boost of your upload speed for a flat fee.
Operators will have to invest in solutions for traffic policy management from vendors such as Camiant, Bridgewater Systems, Sandvine or Juniper. These are the guys that can peer into network activity and then open or close the network valves to let more or less traffic flow to an individual subscriber for an individual service.
Most participants in the Network Neutrality debates are OK with such policy management as long as it is neutral: time-sensitive services like audio and video streaming can be prioritized as long as all such services can get the speed boost, and not just those that the ISP picks and chooses.
On top of this, service providers will require a pricing and charging system, such as Highdeal Transactive, that allows them to invent more creative pricing models for selling bandwidth. Account balances are checked in real-time to ensure credit availability and calculate advice-of-charge, thus eliminating credit risks. The extra revenue gained from the new services offered will help pay for network buildouts, particularly with heavy investments associated with FTTH rollout.
Industries such as utilities seem to have excellently managed the transition from simple metered to new charging structures. In the broadband space, a new approach to architecture is required, which involves integration of policy, charging and network elements. This will allow service providers to:
- create real-time business model management solutions
- offer & charge for new services based on quality of service
- calculate sophisticated service revenues
- manage revenue sharing between multiple partners
- sustain massive transaction volumes across their platforms while rapidly changing their business models to optimise profitability.
The key is to have real-time control over the allocation of network resources, and to charge users according to resources & services used. The tools to do this are already available today - all service providers need to do is apply their creativity to use them to their full extent.