Prepared for the ITU Focus Group on Technologies for Network 2030. Technology plans should be designed to advance the U.N. Sustainable Development Goals. Will the design produce robust infrastructure, everywhere? Will the proposed networks be affordable for all? What will be the requirements for investment and are they realistic for developing nations? What is the impact on competition and the resulting effect on consumer costs?
People more qualified than I will be discussing possibilities like terahertz spectrum and microscopic antennae. What I want to do is put some questions on the table crucial to the cost of access and deploying everywhere. These are questions, not conclusive opinions. The groups making standards very rarely have considered the requirements of the developing world, which include low costs and simple deployment.
Some issues that might arise in coming wireless networks include:
- Possible obstacles to shared, multi-tenant systems. Consider a rural road system which requires many small cells. In many cases, it's unrealistic for three or four wireless networks to build separately. Sharing networks has become common from Canada to England to India. Fewer towers also means less pollution.
One network is cheaper than two. Two is cheaper than the four to seven often required for strong competition.
- Requirements in the standard that drive up costs significantly without benefits for all. While some countries and carriers can afford networks with expensive features, others are held back by the expense. Simpler networks are cheaper to build and to operate. Costly advanced features could be made optional. Huawei has been working on inexpensive units easily deployable. Bringing down the costs is particularly important in less developed areas. In practice, nearly all standards have been controlled by developed countries and CJK. I believe the interests of other countries should come to the heart of the discussion, ideally by stronger participation from around the world.
- Coexistence and efficient coordination with other technologies that share the medium. In particular, Wi-Fi and 4G/5G may share spectrum. There are major efficiencies to use the cloud to reduce interference, often requiring information about the environment. A way for both to communicate would have substantial advantages. Currently, 3GPP and 802.11 have battled over how to keep out of each others' way in 4G & 5G. Both lose capacity. Twelve years should be enough to move from theoretical discussions to field deployments.
- Low energy requirements. The majority of Africa is not covered by electric grids, nor are many sparsely populated rural areas in the developed world. Systems that can work with solar power are ideal. All energy reductions are important and reduce costs.
- Royalties should be reasonable. While there is no international standard, I believe ITU Secretary-General Hamadoun Touré's suggestion of 5-10% for the total retail price is a good guideline. This is particularly relevant to high volume communication devices, such as cell phones. Three years ago, Carlos Slim told me US$50 smartphones would connect the next two billion people. The current royalties demanded would raise the retail cost by at least half. Korea and China have ruled that Qualcomm royalties disrupted crucial markets and Qualcomm has since paid US$1 billion in fines. Qualcomm has since raised the royalty rate from 3.3% to 5%. Clearly, the existing ITU/3GPP system to ensure reasonable royalties is insufficient. The IEEE system that bases royalties on the price of the relevant part (e.g. a 4G chip) rather than the the total price of the cellphone is an interesting approach, but far from sufficient. Standards already consider whether a proposal unreasonably raises costs. High royalties need to be part of that discussion. (They may require a new understanding of competition policy but that needs to be reconsidered.)
- SDN, NFV, and Open Source are proving to reduce network costs. Technology submissions should clearly be able to function in a vendor neutral network. Whether that should be implemented with YANG models, Open Daylight, CORD or whatever, the capability must be there.
- Standards-based networks must be deployable by more than a very few big companies. 3GPP, ATIS, and the companies involved probably breached competition law with 5G NSA. They wanted something giant telcos could deploy two years before the complete 5G standard would be ready. 3GPP hived off the air interface and some mmWave solutions into the 5G NSA release. For practical purposes, 5G NSA could not be deployed except by those with a 4G network in place. In many countries, that meant only 3 or 4 giants. New entrants were essentially impossible.
This is a first draft to get discussion started. I'm asking some of the most qualified people I know for more ideas. Improvements welcome.
Delivering on the 5G promise of increased data rates, and ubiquitous coverages, poses stringent requirements on traditional vertically integrated operators. In particular, telecom operators are expected to massively roll out Small Cells, which requires finding appropriate urban spaces with both backhaul and energy availability. Network sharing becomes essential to unlock those commercial massive deployments. The open access model, or neutral host, will come to play a key role on the deployment of 5G networks, especially in urban scenarios where very dense Small Cell deploymens are required.
In parallel recent trends are paving the way towards the development of new, heterogeneous and distributed cloud paradigms that significantly differ from today’s established cloud model: with edge computing, cloud architectures are pushed all the way to the edge of the network, close to the devices that produce and act on data. We posit that there are two sets of players perfectly poised to take advantage of both trends since they already own the infrastructure needed to build edge deployments: telecommunication providers and municipalities. 5GCity focuses on how common smart city infrastructure (i.e.,small cells and processing power at the very edge of networks) can bring benefit to both players based on resource sharing and end-to-end virtualization, pushing the cloud model to the extreme edge.
5GCity will design, develop, deploy and demonstrate a distributed cloud and radio platform for municipalities and infrastructure owners acting as 5G neutral hosts. 5GCity’s main aim is to build and deploy a common, multi-tenant, open platform that extends the (centralized) cloud model to the extreme edge of the network, with a demonstration in three different cities (Barcelona, Bristol and Lucca). 5GCity will directly impact a large and varied range of actors: (i) telecom providers; (ii) municipalities; and (iii) a number of different vertical sectors utilizing the city infrastructure