How the broadband farmers and cowmen can be friends
The farmer and the cowman should be friends, Oh, the farmer and the cowman should be friends. One man likes to push a plough, the other likes to chase a cow, But that’s no reason why they cain’t be friends.
– Oklahoma! by Rodgers and Hammerstein
A previous post explained why it is inappropriate to naively impose historical common carriage regulations on broadband ISPs. The notion of common carriage simply doesn’t match the nature of the statistically multiplexed broadband resource. However, that doesn’t mean all is well. There are real unresolved problems of fair and equitable access to that resource, and these are a source of legitimate disquiet. Whilst today one side of that ‘neutrality’ conflict is angry and fearful, the cycle will continue, and at some point the roles will switch.
I have been continuing my discussion with Dr Neil Davies about why the ‘telco supply farmers’ and ‘user demand cowmen’ are at loggerheads. Neither side is able to hear and empathise with the other side’s pain, nor are they able to resolve their differences outside of law courts. This process is ultimately damaging to all parties. Everyone needs to make a living from the same shared fertile communications territory. In the long run, the farmers and the cowmen both have to succeed, and not at the expense of the other.
How can they become friends again?
False assumptions cause conflict
Many years ago, I came across a ‘thinking tool’ by Eli Goldratt called the Evaporating Cloud. Goldratt taught that “every problem is a conﬂict, and that conﬂicts arise because we create them by believing at least one erroneous assumption.” Thus any route to resolution of this conflict requires the surfacing of erroneous assumptions of one or both sides.
Both the telco farmers and user cowmen are, in my and many others’ experience, generally people of good intent. They even share the same over-arching goal and need: a large and thriving broadband-enabled digital economy. They are all wholly dependent on the vitality of that ecosystem and their shared communications land. Yet there is bitter division between them. Why so?
The conflict that exists between them is due to the way that they have framed the nature of broadband as a resource, and their erroneous assumptions about its capabilities and costs. The telcos are not realising the full potential of the broadband networks they creating, and users are hoping broadband will do things that it can’t do.
Broadband networks are not ‘pipes’
These errors all have their root in a common failure to engage in three simple facts about broadband networks:
- They are a shared, and thus contended, resource.
- That contention cannot be destroyed, only re-allocated.
- Every choice over re-allocating that contention is always a trade
That implies networks are trading spaces (and at all timescales). This is because when you grant access for a shared resource to one source of demand, you implicitly deny it to everyone else. There are no exceptions! Even a lowly FIFO tail-drop queue is doing trades, whether or not consciously observed and managed.
Networks are more like derivatives exchanges than ‘pipes’: there are ‘buyers’ and ‘sellers’ of present and future contention, based on an underlying physical ‘information translocation’ resource.
The ‘neutrality’ issue is at its heart about control over this trading space to access the resource. The conflict gets its poisonous energy because both sides reason about broadband through analogies and metaphors, pejoratively couching the other side’s framing as manifestly wrong. Yet neither side has engaged with the true nature of the system under discussion!
So what is each side’s relationship to the trading space, and the false assumptions they have?
The telco farmers’ beliefs
The telcos have tried (and often failed) to use traffic management and QoS mechanisms to help them make better resource trades. Their ability to configure these mechanisms to meet a particular performance or cost outcome is relatively weak. Indeed, they construct and run multiple (single class of service) overlay networks, delivering voice, video and Internet access. This prevents any dynamic trades between these statically allocated resources.
They also fear political or regulatory punishment for ‘unreasonable’ traffic shaping that exploits trades, as they become lightning rods for the inherent tensions in the broadband model. The idea of creating new advanced assured services that exploit resource trades is seen as highly controversial, and may indeed be actively discouraged by some regulators.
As a result, they feel there is no alternative to getting a return on their investment other than good old fashioned way: exploit any termination monopoly available.
What they don’t realise is that the trading space is far richer than their networking textbooks tell them, which only hinted at the trades on offer. The QoS mechanisms their vendors supply only expose them to a small proportion of the possible ‘good’ trades.
Indeed, much of the focus in network research and development has, erroneously, been on making contention disappear. This defeats the purpose of packet networking, which is to enable more intense resource utilisation through sharing. It is the equivalent of building roads to ensure there is always capacity to avoid a jam. As a result, they are only squeezing out a fraction of the value-creating capability of their networks. Their capital is grossly under-utilised because of the way they have constructed and configured the trades.
Meanwhile, users are frustrated with the poor performance of important applications, and many would be delighted to get new services that deliver experiences they can’t get using commodity Internet access. There just isn’t any appropriately formed wholesale market yet where the access providers, CDNs and applications can create the necessary supply chain.
The user cowmen beliefs
The users and their policy representatives often have a naïve view of the network trading space. There is an entitlement culture that believes that users should receive the experience of the wide open country where no other cowmen (and their associated livestock movements) will ever get in their way. Any failure to deliver is down to mean and evil telcos ‘degrading’ their experience and failing to invest in the network to create artificial scarcity.
The user cowmen believe that any kind of ‘trades’ will be exploited by the telco farmers to their disadvantage. After all, the telco farmers have a long and inglorious history of exploitative monopolistic behaviour: they erect fences to enclose the land, and assert ownership over unused territory, forcing it to lie fallow even there are others who could eke out a meagre existence upon it.
Therefore the user cowmen endlessly demand more capacity to reduce contention effects, for which they are not prepared to pay any premium. They too seek to defeat the point of packet networking, which is to intensify the use of a shared resource. Telcos find this economically unsustainable, and thus ultimately must refuse to play the game.
As a result, users they feel there is no alternative to getting a decent experience other than good old fashioned way: call in the regulator to prevent the telco from exploiting its natural termination monopoly to fund the users’ own unsustainable demands.
What they don’t realise is that there are ways of constructing networks that can exploit the trades, but don’t turn the telcos into exploitative toll keepers. Control can be kept with the users, individually and collectively.
The regulator is stuck in the middle
The regulator is left in an impossible situation. On the one hand, they are being tasked to promote the development of broadband. That means raising capital, for which telco farmers must get a good return to ensure the land gets improved. You can only get a good return if you deliver valuable experiences (which means good performance and dependability), and have low costs.
On the other hand, they are being told by users and broadband campaigners that they don’t want to have the telco involved in any way in helping to create those good experiences and manage the costs.
Regulators aren’t helped by the lawyer and academic classes clinging to a fantasy that FIFO queues represent some kind of ‘neutral’ mechanism, and that all packets were created equal in some kind of Rawlsian system of information justice. (Don’t even ask about their cost models.)
The real issues
To come to some kind of reconciliation, everyone is going to have to align with a non-negotiable broadband reality: there is a shared resource, and a trading space for its allocation. There is only one land, and they both have to extract their livelihoods from it.
Both sides want good application outcomes (revenue to telcos, good experiences to users), and high resource use (makes it profitable to telcos, and lowers prices to users). The only way to achieve this is to make lots of good resource trades.
The issues that should be on the table are:
- What resource trading capability should be created?
- Who controls the trades?
- How do they get offered commercially?
At that point we’re in a rational universe of discourse, and reasonable people (and even their regulatory lawyers) can negotiate based on a shared and reality-based understanding of the resource in question.
How the farmer and cowman can be friends
We can see that there is only one long-term way of resolving this conflict: to understand and embrace the trading space, not to try to make contention disappear. This means a polyservice network that properly exploits the trades. This should not come as a shock, since every Internet Protocol packet has bits reserved in its header in anticipation of such a thing. It’s just up until now some basic misconceptions about the nature of the network as a trading space have gotten in the way.
One you have such a network, which can offer differential levels of quality and assurance, you can then begin to talk about terms of access. What you care about is the ability to contract to some kind of outcome for service assurance in a non-discriminatory manner.
‘Network neutrality’ is a category error, because it misattributes the equality to packets. What matters is equality between network users, and their fair access to whatever services the network offers that exploit the trading space.
After all, if broadband networks are trading spaces by their nature, then they are also naturally marketplaces. The telco farmers can produce crops of high performance, which gets sold to the user cowmen to feed their application stock; those animals in return help fertilise the land with dirty old cash. Everyone gets to eat a feast of great customer experience beef at the end of the day, washed down with profit-infused beer brewed from the farmers’ crops.
Once that happens, everyone can be friends again.
Martin Geddes (with credit for the original ideas to Neil Davies)
To keep up to date with the latest fresh thinking on telecommunication, please sign up for the Geddes newsletter