Hiding OSI layers leading to policy failures: Net Neutrality, Encrypted Media, ...

I've regularly written about the OSI model for digital communications.  I'm increasingly concerned that policy discussions that hide these layers are leading to policy failure.  I see this in the odd rhetoric coming out of the USA on network neutrality, which is the same country that originated the policy failures around technological protection measures.

When I learned about digital networks in the late 1980s this included the Open Systems Interconnection (OSI) model which describes the different interoperable layers that  digital communications enables.  This is in contrast with analog networks which were purpose built and inflexible.  Each layer in a digital network is built upon other layers, starting with the physical layer that describes specific electronics all the way up to the application layer which is the layer closest to the user.

When discussing communications technology people often make analogies to transportation technology, since most people have a greater understanding of transportation technology.   The problem is an poor and inflexible analogy has become dominant.

A comparison is often made to shipping via boats and rail, where a primary policy is common carriage.  This analogy suggests digital networks only have one layer much like the older analog networks, wiping out the flexibility both in terms of technology and policy which digital networks enable.  This flattening of layers also causes policy confusions which wouldn't happen if the layers were exposed through a better transportation analogy.

A layered model for road transportation

In 1994 the federal government formed the Information Highway Advisory Council (IHAC). Discussing roads and highways is an appropriate analogy to communications technology as it exposed the layers and complexity of the network, even though road transportation is still less flexible than digital communications networks.

A simplification of layers built on road infrastructure might be:

• Road infrastructure.  This is comparable to the physical network layers.
• Vehicles run "Over The Top" of those roads.  This is comparable to physical devices connected to the communications network.
• Drivers control the vehicles.  This would be comparable to software authors, where software is the instructions that drive digital devices. (Note: It is software that differentiates between TCP/IP and other networking protocols.  ISP's are businesses that run their own devices and provide transport of packets encapsulated within TCP/IP.)
• Passengers and parcels which would be placed in/on the vehicles for transport.  This is comparable to the applications which use the network (two way or one-way audio/video/text/etc communication) 

With transportation the roads are a mixture of municipal, provincial and federal management.  Private roads including driveways connect to publicly managed infrastructure.   While publicly owned vehicles exist, private (corporate and individual) vehicle ownership far exceeds public.  Individual citizens are allowed (in many ways actively encouraged) to personally own and drive vehicles.

If we use this road transportation analogy to go through various policy discussions the failures becomes more obvious.

Technological Protection Measures

While Canada formed IHAC, the USA formed the National Information Infrastructure working groups. Bruce A. Lehman chaired the Working Group on Intellectual Property Rights which during 1995 came up with a disastrous concept:  if it was possible for vehicles to be used to transport something illegally, then private citizens should not be allowed to drive vehicles or choose drivers.  Since immediately outlawing private drivers would have been too controversial to pass, a mixture of law and market forces would be used to indirectly achieve the goal.

• Vehicle manufacturers would be granted the right to impose drivers, and it would be made illegal for the vehicle owner to fire that driver and chose their own. While a private citizen might be allowed to "own" a vehicle, they are not given the keys to the locks and it is made illegal for the owner to change the locks.
• Destinations would be given the legally protected right to deny access to any person who did not provide proof that they arrived using an "authorized" vehicle with the manufacturer imposed driver.  The ability to access these destinations would serve as a market force to impose manufacturer determined drivers onto the majority of the population.

Mr. Lehman and his supporters may claim they were only trying to reduce unlawful activity, but it should be obvious that the harm to the economy and society as a whole of this type of policy greatly outweighs the alleged harm he was claiming to reduce.

This is the essence of the policy which Lehman tabled, was policy laundered through WIPO in 1996 when the 1995 bill didn't pass within the USA, and which later became the controversial part of the DMCA in 1998.  Canada included this harmful policy in Bill C-11 which inappropriately provided legal protection for "access controls" (IE: ties between content and specific devices/software, and non-owner locks on devices), even though this was not required by the WIPO treaties.  An even worse variation of this harmful policy was included in the TPP, and I will be surprised if the USA doesn't try to push this as part of the NAFTA renegotiation.

Network Neutrality

Imagine a country where a tiny subset of retail outlets owned all the roads. Governments and lobbiests would claim that there was "competition" in road infrastructure if home-owners in a specific city were able to choose between connecting their driveway to the Walmart roads or the Loblaws roads, where these roads favoured in subtle and not-so-subtle ways the ability of people to access some destinations over other destinations.

Companies like DHL, FedEx and UPS might be allowed to exist, but would be disparagingly called a "wholesale" market of the services of Walmart and Loblaws, rather than recognizing shipping as a different type of service than vertically integrated road owners which might also own their own vehicles and do their own shipping.  (Digital Example: Companies like Techsavvy are claimed by the CRTC to be part of a "wholesale" market, even though their TCP/IP routing service is built on top of the same physical infrastructure).

While all surface transportation related services are built "over the top" of the physical layers, the "over the top" terminology would be abused to refer only to competing services.  Even if you wanted to buy the identical item from Metro or Loblaws, the purchasing from Metro would be called "over the top" while the purchase from Loblaws would not.  The nearly identical service offered by Loblaws/Walmart would be regulated differently (or prohibited) if offered by a competitor (Digital Example: Bell's IPTV service branded as FibeTV is regulated as a cable service, even though our Copyright Act explicitly disallows this type of new media retransmission without separate permission/payment).

In a road neutrality debate the US Department of Transport chairman might claim the fact that so many people get entertained at Cineplex theaters is somehow "proof" that road neutrality already doesn't exist, so what tiny amount of minimal regulation currently exists should be repealed.



This may read as utter nonsense that no government would ever allow, but this is essentially the situation we are in today with communications networks. Incorrectly regulated convergence allowed the incumbent phone and cable companies to gain all the benefits of the OSI layered digital networking.  It now doesn't matter which physical connection (coax or twisted pair) comes into the home or business, the same services can be built on top such as two-way voice (previously called telephone), one-way video (previously called Cable TV), and many other applications.  Just as Walmart now sells food and Loblaws sells clothing, "phone" companies sell streaming video and "cable" companies sell two-way voice.

Policy Solutions

While common carriage has a place in the policy mix, it needs to be thought of as one small policy lever among many -- and only applied to services at specific layers of the communications stack.

As we have nearly always done with transportation technology, structural separation of communications technology is required.  I've come to the conclusion over recent decades that anything less than structural separation will be as effective as deck-chair rearranging on the Titanic.

The model we use for roads was created at a time when it was understood that roads were the infrastructure upon which much of the economy and society was built.  It is entirely inappropriate for the ownership or control of the similarly critical infrastructure for the new economy to be in the hands of a small group of private sector entities.   As with roads the different levels of government might hire private sector contractors to do much of the work, but the ownership and control must rest with the public sector.

We need open competition in the other layers.  The need for foreign ownership rules only apply to the physical infrastructure, which I've already suggested should be managed by the public sector. Other layers are already recognized with transportation to not need those restrictions: while there are Canadian automobile manufacturers, people are allowed to purchase foreign designed and manufactured vehicles.  The same should apply to digital communications products and services that run "over the top" of the government managed infrastructure.

Private citizens must have the right to own their own vehicles, and choose their own drivers for these vehicles including being their own driver if they have the skills.  With digital technology the equivalent is the right to own their own devices, to choose their own software, and to author their own software if they have those skills.   Laws which legally protect non-owner locks on devices, or allow content providers to impose specific device manufactures/software, should be repealed immediately.

The Ministry of Communications needs to be restored to federally mirror for communications what the federal Ministry of Transport handles for transportation.  This is a ministry that was abolished in 1995, at the time when convergence was being mismanaged.  The CRTC is currently inappropriately administered through the Minister of Heritage, who is in a conflict of interest with specific types of communication entities.  The pre-convergence Telecommunications and Broadcasting Acts are in critical need of modernization or replacement.

We as a society have always subsidized arts and culture, which are not always able to be adequately privately funded. While public arts funding should clearly exist for works created to be distributed by communications networks (such as scripted video programming), this should be done through direct accountable public funding and not through cross-subsidies between layers of the communications network.  There are too many ways to get cross-subsidies wrong and for governments to be manipulated: much of the current discussion around a so-called "Netflix tax" is a dishonest misinformation campaign initiated by vertically integrated companies like Bell, Telus and Rogers (often through the TV stations/studios and other media they own).

It should be understood that incumbent vendors will not be happy with any policy corrections. The increasingly extreme policy proposals coming from Bell, Telus and Rogers are to be expected as they are dealing with an existential crisis.  Required structural separation and free market competition would put many of these outmoded companies out of business.  This should be understood as a good thing, not something to minimize or delay, as structural separation will lead to a more innovative economy and society.  It really is a win-win scenario for nearly everyone, and is as critical to our future as governments building and maintaining the road infrastructure has been for the industrial economy.

---

Google Doc version (Which you can print or download a PDF from).