A little history …

I often hear the likes of Google calling for “net neutrality” and ‘flatrate pricing’, yet their business models of true spot-pricing for web advertising really makes me think they are being just a little hypocritical. So how can IP carriers support “net neutrality” and preserve market-pricing at the same time?

Well, the following is a small excerpt, originally written for TotalTelecom, that attempts to address some industry confusion around the interpretation and implementation of “Net Neutrality”. In effect this post summarises material from an earlier article I wrote, which was published in the IEC Annual Review of Broadband Communications Vol 3, (Communications Vol. 60), on NGN’s, technology choices and their subsequent business model impacts. That article, “Which Takes Precedence: Your New NGN or Your Current Business Model“, examined just one of the many key business implications that results as carrier networks migrate away from the multiple overlay networks of yesteryear towards the single, fully integrated all-IP network of tomorrow.

Excerpt :

As carriers around the world move to modernise their existing, multiple, overlayed network infrastructures; they will increasingly be faced with this problem–how to differentiate services delivered over a (now) common network? This differentiation needs to be more than just relative, it must include expressed characteristics in order to permit differentiated pricing. And differentiated pricing (or market based pricing) is absolutely vital to ensure the longterm performance and profitability of these next-gen networks and their operators. Interestingly enough, market based pricing also ensures a optimally competitive marketplace where both end-user consumers and service/content suppliers win as each can ‘buy’ in at any level they feel meets their expectations. This allows for universal appeal, which is a very different concept to universal acceptance.

It is clear that carriers of the future will need to strongly consider offering a number of clearly differentiated (albeit still common) path pipes rather than just a single transparent bitstream pipe concept for connectivity services. These pipes need to be able to offer QoS differentiation that ranges from ‘least effort’ style services, through QoS that is ‘business grade’, then QoS that is ‘assured throughput’ and finally QoS that is ‘expedited performance’. I believe this represents the minimum (and sufficient) set of services. Although it is possible with increasing management complexity to offer additional services in-between those four. For each QoS class, carriers should also differentiate between a committed information rate (packets carried end-to-end with 100% or close reliability, excluding outages, and to which paths are capacity dimensioned) and an excess information rate (which must be offered and desirable, and is carried either as best effort or can be statistically guaranteed some level of delivery probability through appropriate burst and buffer dimensioning).

Enforcing these classes to ensure that they are not only relative in priority but actually perform differentially under various network load and utilisation conditions requires that all traffic from all sources of a common class eventually aggregate to share a common forwarding block (ie. same network path, same interface queue and same egress scheduler). This ensures that all users of a given class commonly share all the network performance characteristics of that class at a given instant in time. This criteria protects the carriers ability to then differentially price such classes, safe in the knowledge that network users will be both unable to reliably arbitrage the system and/or unfairly leverage the network services beyond their purchased QoS performance.

Armed with such a network, carriers can then immediately begin selling services into the market at whatever price points customers are comfortable to purchase at (Economy, Business, First, Exclusive to use an airline analogy). The concept of net-neutrality is also preserved by defaulting all traffic to economy UNLESS the users or providers have specifically indicated a willingness to pay for higher performance carriage in which case their traffic can be up-classed and the appropriate party charged for the privilege. With this architecture the world of telecommunications can move on, having successfully avoided a ‘tragedy of the commons‘ and simultaneously enabled an ‘open, competitive market place’. (For the record, the peak available bandwidth within a providers aggregation or interconnect network is a limited and finite resource during the peak utilisation periods seen as part of the natural diurnal cycles. Although this resource can be increased arbitrarily, doing so comes at a cost to the provider and must be passed on to subscribers in some way to be long-term sustainable).

Finally, even more convenient, if carriers adopt the four class model described above and leverage the ‘in-profile’/’out-of-profile’ markings for each class, then they are able to negotiate interconnect and handover points using common layer-2 network technologies (MPLS and/or Ethernet) that preserve the intended end-to-end performance characteristics without having to modify or overwrite the end-users IP DSCP headers. In effect, they can offer total end-to-end transparency (key for network neutrality) whilst preserving the ability to differentiate end-to end performance across (feasibly) multiple network operators. And last but not least, if network operators can support standard interfaces for end-users and their content providers (the over-the-top players) to signal willingness to pay for services and applications, not just to themselves but also to/from the operators, then we have a win-win-win scenario. It is not good enough to try to lock QoS to applications, ie. low priority to Youtube and high priority to Heart-rate-monitors, as there are occasions where my heart-rate-monitor is just a background thing that happens for insurance reasons whilst a particular Youtube video (commercial release of content) I may have paid to receive as an early release. Flexibility in service offerings is the key for the future.