6G - The most important question!

If we do not answer this one fundamental question correctly then we will standardize functionality that will never see commercial success nor create the required returns to fund future success.

The question:

Are we designing for 6G dedicated devices or multi-network devices?

This is the most important question to answer. Before we do anything else. It will define what should be included for consideration in the scope of the standard and what should be left outside that process and in the hands of the market.

I believe we are currently answering it as if it is still 1998. We are writing the standards using the assumption of ubiquitous connectivity being provided by one version of one standard. This has nothing to do with reality. We are living in an increasingly heterogeneous network world, one that is tremendously more powerful through the diversity of alternative solutions, and the different system and economic characteristics they bring.

We behave as if our biggest device customer, the smartphone, never adopted Wi-Fi as a partner network. If we do not embrace the opportunities of reality that surround us then we deserve to go the same way as other very good technologies that failed to keep up with the times.

There are devices that will be tethered to one technology but we have to understand the market size and commercial returns attached. These are not billion device homogeneous markets with high ARPU. We have to decide the best way to address what needs and only standardize what is materially valuable to the largest common underlying strata. To standardize the world based on a small part of the addressable market introduces cost, complexity, and delay on the masses, in service of the niche few. At the end of this article I refer to these types of solutions as "Specialized Coverage" opportunities. I do not believe they all look the same.

Will these devices be tied to one variant of one technology? Are we ensuring backward compatibility of these incredibly complicated higher level functions? A car requires 30 years of guarantees, will it implement the latest invention in a standard and trust it will exist, be deployed and be supported?

In my opinion it is time for a hard reset.

And that starts with blowing up the biggest fallacy of all. The one network fallacy.

History since 1997

"The Rise of the Stupid Network was originally written during a single long weekend in May 1997 by David S. Isenberg while he was an employee of AT&T Labs Research. The focus of Isenberg's article is the development of a network that is smart at it's endpoints (where users determine it's activity). This is in contrast to so-called smart networks built by phone companies which are composed of stupid endpoints (phones) and smart networks that handle all the call control. Isenberg's main point is that the Internet, a stupid network, is creating new value systems by transitioning the network paradigm from smart to stupid." [ref]

David Isenberg left AT&T in January 1998. Shortly after that, AT&T asked Isenberg to remove The Rise of the Stupid Network from the isen.com website. Isenberg wanted to make AT&T happy so he took it off. (It didn't work; AT&T is still not happy.) [ref]

This was written in May 1997. This is before Wi-Fi. This is before mobile broadband. This is before everything we now take for granted. And he predicted what was to be.

The name, "stupid network", did a disservice to the whole concept, in the same way "dumb pipe" repeated the mistake many years later. Nobody wants to build something stupid or something dumb and a network that connects the planet, at exponentially decreasing costs and exponentially increasing capacity, is far from dumb.

The network needs to be good at what the network needs to be good at and the device needs to be good at what the device needs to be good at, in a world of many devices and many networks. We need to understand the different between the two.

We need to listen to this message, embrace it, and implement standards that simplify rather than complicate, that remove technology debt rather than add to it, and that focus on being the best network that enables operations with other networks seamlessly.

The Device (e.g. Smartphone)

As of 2023, smartphone subscriptions account for approximately 82% of all mobile subscriptions worldwide (total 8.46 billion mobile subscriptions). By 2029, smartphone subscriptions are forecast to reach 8.06 billion, making up about 87.5% of the projected 9.21 billion total mobile subscriptions [ref].

Percentage of topline revenue attributed to smartphone subscriptions is even more polarized given (taking USA as an example) non smartphone (IoT) plans range from about $5-$10 per month, compared to smartphone plans of $50-$60 per month.

Smartphones became multi-network by including Wi-Fi as standard, starting in 2004 [ref].

For a lesson in the progress of Wi-Fi network parity I recommend the following excellent paper by Mark Grayson .

For fiber first markets (alternative being mobile first markets) Wi-Fi networks carry upwards of 70% of all smartphone traffic. By volume, Wi-Fi is the dominant network rather than the supporting act.

In addition the largest device platforms are abstracting the developer experience away from understanding which network a device is connected on, and the evolution of internet protocols are creating session continuity on top of switching networks and multi-path transmission to maximize performance across all possible paths.

In 2023, the Internet is continuing to transition. Not only is over 90% of Internet traffic encrypted, in certain regions of the world we observe that nearly 50% of the traffic has transitioned from regular TCP to HTTP3 transported over User Datagram Protocol (UDP)- based QUIC[10]. Critically, instead of having to mask different paths, the QUIC transport protocol supports native connection migration. Existing connections continue to operate as devices change their endpoint IP addresses when they switch between different networks.

The device ecosystem is looking to meet the needs of their application providers by delivering frameworks that enable applications to configure how multiple paths should be employed, enabling applications developers to easily benefit from the HTTP3 connection migration capability.

The device is no longer connected to a network, it is connected to an abstract concept of connectivity. Creating technology that assumes anything else will not intersect with where the devices are moving.

The smartphone device IS why we build mobile networks. We need to take care of this device experience as the number 1 priority for the immediate future. It pays for adjacent exploration. We need to take care of it by providing the best network behaviour and cost structures, that can be integrated into an operation of many networks where all networks can be easily treated as equal citizens.

Internet of Thing Devices

Are these devices dedicated to one network?

The trend is increasingly no. And if they are, they do not want to be tied to one generation that needs replacing with the next.

Most modern IoT solutions are designed to work across multiple networks rather than being tied to a single one.

Multi-Network Capabilities

1. Global Coverage: Many IoT connectivity providers offer solutions that work across multiple networks globally.
2. Multi-IMSI Technology: This allows devices to connect to multiple networks using a single SIM card.
3. eSIM and eUICC Technology: These technologies allow for flexible network selection and switching.

While there may still be some IoT solutions dedicated to a single network for specific use cases or in certain regions, the majority of modern IoT connectivity solutions appear to be designed for multi-network compatibility. This approach allows for greater flexibility, scalability, and global reach in IoT deployments.

The exception is when coverage is required for specialized interests, for example mines and ports that have deployed dedicated private networks.

A need for a new Framing

Dean Bubley has suggested a revisit to the IMT 2030 6G hexagon.

This is a good approach since it aligns with the current operational approach.

However it struggles when applied to the above multiple network reality, and if not aligned the functionality will be hidden inside the overlay layer above. I think the exercise that Dean suggests is a valuable one to execute for vested parties.

A more radical alternative

We move the hexagon outside the standard and allow the market to decide what makes sense. Inside the standard we change the goal to creating maximal network usefulness (again a Dean Bubley construct) with the standardized capabilities delivered by the technology aligning with making the network as good and commercially viable as possible (see commercial differentiators) rather than defining use case prediction within the standard process.

The network supports all uses cases through being a good network supporting all required characteristics and enabling interworking with all other network types.

Introducing each layer from right to left...

Maximal Usefulness

All models have a goal. The goal of the 6G model is to maximize usefulness across all coverage interests contained in the model.

Coverage Interests

Traditionally there has been one commercial customer stakeholder type - the nationwide operator, who has licensed spectrum across a country and committed to providing coverage to all parts of the country as part of the license. This has never completed due to economic challenges in less densely populated areas and also it is incomplete in urban areas where "not spots" of service still exist.

This view of the world was created before alternative mainstream network technologies such as Wi-Fi (and now direct to device cellular) networks were an alternative. Yet the whole mechanism of the standards approach is still to assume ubiquitous coverage being offered by one homogeneous network 3GPP based technology, and to standardize high level functionality that requires complete coverage.

This is a false premise and leads to higher level functionality becoming a cost and a complexity. This leads to bloat, expense, and time.

A better approach is to define coverage areas of interest where an operator may choose to offer service in one or more area types and then apply the most appropriate technology to each.

In the future there will be less a coverage problem than a seamless accessibility problem. For example Wi-Fi for indoor, moving to macro for urban moving to cellular for rural, and a combination Wi-Fi, cellular access network where both are needed.

It is not the job of mobile network operators to monetize the 3gpp standard. It is the job of the standard to help mobile network operators monetize the market opportunity, as one technology choice of many, in a total service operation.

The other insight is that all coverage interests, with the exception of specialist coverage, support both work and living. The concept of being in a specific location for non-specialized work has gone. The concept of not living where you work is gone.

The only important definition is whether the location is single tenant or shared.

Each coverage interest has its own definition of ubiquitous coverage to create maximal usefulness and the deployment is sized to meet the economic minimum to achieve that goal.

This framing means nationwide ubiquitous coverage only applies to one coverage interest - the nationwide operator. And there needs to be revisiting of appropriate and creative incentives to enable this constituent to achieve appropriate ubiquitous coverage as required by a government's responsibility to provide social good for all residents.

Commercial Differentiators

The commercial differentiators should have equivalents in any other network technology.

Energy efficiency: Minimize the energy costs, energy wastage existing.

Data Telemetry: Standardize all observational data that can then be repackaged into operational autonomy, metering, higher value second order opportunities (behavioural guarantees, fraud, ...)

Security/Privacy: Ensure maximum instrumentation of trust and integrity.

Capacity, Performance: All metrics with respect to number of devices, total bandwidth, latency, jitter, ...

Operational Automation: A focus on operational simplification to reduce cost and speed of operation, from planning, building, operating, through life cycle managing.

Enabling Domains

These are the scopes of interest where architecture, and requirements can be applied, to best meet the defined customers to the right.

Spectrum: How can spectrum utilization, access and sharing be maximized across all interested parties.

Access Networks: Where are access technologies are treated as first class citizens in the process, without preference for one over the other.

Core Networks: The decoupling of access network to core network to enable open control from different operating DNA companies.

Authentication: The decoupling of authentication preference to enable federation and seamless access of all network types and operating entities.

Devices: What devices support what technologies for what applications.

Identified Stakeholders

I have identified 3 main stakeholder types.

Governments to ensure country competitiveness, equity and governance - for both economic benefit and national security reasons.

Operators who are the commercial operating entities that use the standard content to generate economic returns that creates a growth industry.

Technology providers that enable the standards to be translated into goods and services to power the operators.

None of these three stakeholder roles are independant and can be combined however.

In conclusion

This model is not intended to offer a complete or correct perspective of a 6G alternative.

It is intended to provoke a different thought process to highlight the gaps of today and to allow a discussion on what an alternative approach and model could look like.

I hope this starts discussion on what is missing, what is wrong, and what is important to remember.

I hope we also acknowledge this is a complex task and pay credit to the standards work done to date, to get us this far.