Our 7 bold predictions about NET 2.0: 15 years later

By Zheng “Bruce” Li and Jan E. Berglund

In this follow-up to the original “NET 2.0: a critical look at the mobile communication business”(circa 2005), we will assess the validity of the seven original predictions, examine the missing signals and macro trends, as well as propose an updated innovation strategy for networking business in the next 15 years.

Seven original predictions analyzed

1. From centralized to decentralized content generation and distribution

This thesis is half true: content generation is much more decentralized, while content distribution is still largely centralized.

Quality content still rules, HBO Game of Thrones still pulls in one of the largest viewership. But that quality does not have to come from a big studio any more. Authentic, passionate and original content from individual vloggers can enjoy a significant share of their particular niche, and become ever more valuable to advertisers and direct to consumer brands. Case in point: an 8 year old kid on youtube earns over $26M a year by unboxing and reviewing toys.

The main caveat is that centralized platforms like YouTube and Twitch are controlling the lion’s share of the user generated content. Content creators, except the top 1-5%, have little bargaining power in dealing with these monolithic platforms. Steemit and DLive are two of the decentralized content creation and distribution platforms that have garnered a decent number of active users, but it still pales in comparison to any of the centralized platforms.

More streaming contenders crop up almost every week, adding to a long list of existing streaming services like Netflix, Amazon Prime Video, YouTube TV, Hulu, HBO Now. So in this sense, there is slightly more diversity and thus competition in the content distribution channels. 

• 11/1/2019: Apple TV+
• 11/12/2019: Disney+
• 11/19/2019: Google Stadia
• Spring 2020: Quibi
• April 2020: NBC Peacock
• May 2020: HBO Max

2. From subscription based access to advertising and services subsidized access

This model only found its way into a niche of the market.

Most of the communication services we use, be it mobile or fixed home broadband access, are paid via traditional subscription fees to the Mobile Network Operators (MNO) and Fixed Network Operators (FNO). From a MNO/FNO perspective this can be considered a failure to monetize beyond the direct consumer model. While we saw efforts from several attempts of the network operators to try to participate in the digital economy it has largely been in vain. Part of this is due to the geographical constraints which limited user base, and also to some extent data protection rights limiting what can be addressed.

The notable exception for this is wireless access in public venues such as coffee shops and supermarkets. Here the operators still operate the service but it is subsidized by the venue who captured the digital value of users while subsidizing the access.   

3. From industry driven to community driven infrastructure evolution

This is a slow process. While there is no clear repeatable pattern on a large scale, there are a few processes here and there in play with different touch. In the original thesis we pointed to a demand from underserved groups serving themselves as they had been let down by the MNO/FNOs

1. Some of these underserved groups have been served not by taking their own initiative, but through regulation stepping in on their behalf. Regulators bundled coverage requirements when new spectrum assets have been released, e.g. in the UK and Germany,  in order to bridge the digital divide. Much of Europe has dedicated funding subsidizing infrastructure builds for  broadband services, to equalize the access also for financially non-viable areas.
2. Other local efforts such as city networks where a concession holder is allowed to build the fibre network in a specific area to serve all users in the area, reducing the need to dig the streets more than once.
3. The rise of neutral host providers, where landlords take action to solve the indoor coverage to attract tenants (business or residential). A new breed of players, neutral hosts provides, serve the landlords with technical know-how and relationship to the MNOs. Examples are Stratto OpenCell, WIG, CellNex.
4. Community driven efforts such as Ch4LK Mobile, using unused MNO spectrum to create local coverage and then connecting the MNOs to that coverage.

There are a few conclusions from the above. While the need for communities to have service is clear, it seems the path of least resistance has been through lobbying rather than taking matters into their own hands. It is also clear that the MNO/FNO’s still serve a purpose even after the alternative coverage has been built. The community can benefit from the non-technical part of the service such as help desks, package bundles, and billing. 

4. From service provider decided coverage to user contributed coverage

This is partially true.

We have also seen the Fon opt-in schemes allowed and promoted by many leading operators in Europe and Japan such as BT, SFR, Telecom and Softbank now serving 23 million access points, where users can take part allowing access in return for free access provided by other users. In the US, Comcast Xfinity WiFi provides a similar shared WiFi service to augment their mobile MVNO service.

More free standing models, non-operator sponsored, have not had success. It is partly due to service limitations for the end user, and partly due to the business model prevalent among FNOs where modem/access point are bundled with the service and hence don’t feature that model of access. 

The rise of neutral host providers, where landlords take action to solve the indoor coverage to attract tenants (business or residential) is a clear path in this direction. For landlords with high value properties this is a no brainer. Good office coverage is now a must, but in many cases the MNOs do not see the direct value of this and thus leave the investment to the landlord.

5. From licensed spectrum to open spectrum and cognitive radio

This thesis is only partially realized, albeit at much slower speed than we would like.

The MNO’s are still favoring operator controlled assets, i.e. licensed spectrum. This is one of the few but significant market entry barriers that remain. Regulators from time to time try to look at opening up spectrum for other interested parties, both for new use cases in private networks but also to allow underutilized assets to be used more freely.  Today we are still far away from cognitive radio. And where spectrum are more freely shared such as E-band, the licensing is either case by case or “uncoordinated” with restrictions on causing interference and giving way to whoever used the spectrum first in that regional space. 

The battle continues for the spectrum where the Mobile networks fights almost all other groups for access to more dedicated spectrum. Given the value contribution by the MNO, this may not be a totally unfair position but it risks stifling innovation and alternative use.

Looking at the state of the business in large parts of the world such as China, Europe, Middle East, Japan and Korea, there is a question if more mobile broadband competition is needed. The services provided by the MNOs are reasonably affordable. Data quotas are not directly hampering media consumption. Coverage, while not perfect, has improved quite a lot since our last piece was written. As noted in some of the other thesis above, being part of a larger network/service has some benefits to the end user. And in those markets the MNOs are not unreasonable in their spectrum hoarding. There are models under which the MNOs’ spectrum pool can be shared under some conditions.

WiFi, the champion of unlicensed technology, is powering ahead with a rapid evolution reaching gigabit speed in the latest iteration. It has built a sizable industry of multiple billion dollars and predicted to grow to over 60 billion by 2024. Chris DePuy, Founder and Technology Analyst of 650 Group, wrote to us:

"We expect the installed base in WLAN devices to grow from 7.8B in 2014 to 62.0B in 2025.  The installed base was zero in 1999 and has grown exponentially since that point. We see similar growth rates for Bluetooth, Zigbee, LoRa also.  Also, 802.11ah is coming to market next year in the <1 GHz spectrums – that’ll grow, too, and few will call it WiFi.

As for cellular, we see something similar happening, and I discussed this at the most recent WiFiNOW conference in London – that as unlicensed becomes available for cellular, it will be adopted as well.  In the US market 5 GHz is open for cellular use (ATT and T-Mobile are using it now as LAA and the MulteFire Alliance is certifying devices in that spectrum in 2020). We have “lightly licensed” spectrum in the US called CBRS, which is not a zero-cost model, but offers what is kind of similar to what Germany and UK are offering as low-cost corporate cellular).  I expect MulteFire to certify additional spectrums including 2.4 GHz and some Japanese spectrums."

So, with WiFi, Bluetooth, Zigbee, LoRa, 802.11ah and cellular all taking advantage of unlicensed (or cheap-licensed), I expect the bottle to be uncorked, unleashing significant new applications and uses that were not here before.

Shared license technology such as 3.5GHz CBRS and SAS are enabling potential newcomers to this market, and notably:

• Cable MSO: Comcast and Charter
• Google Fi (previously Project Fi)

CBRS mobile device support is improving: Google Pixel 3 and 4, iPhone 11, Samsung Galaxy S10 and Note 10, LG G8 and V50 , OnePlus 7 support CBRS bands and have the CBRS Alliance OnGo certification. (Ref: https://www.cbrsalliance.org/certification/)

It clearly shows the power of open innovation enabled by unlicensed or shared spectrum, which in turn spurs collateral changes in the cellular industry for the better interest of consumers.

6. From bundled to decoupled access and service provider

This has been a slam dunk. The MNO/FNOs have largely lost out on this game. Look at the success of Netflix, Amazon prime Video, Spotify, WhatsApp, and Zoom video conference. 

The independent over the top (OTT) service providers being able to effectively work horizontally across the globe has proven to be a winning concept. While the operators still have strong hold onto the access, be it wireless or fixed, they have largely failed to capture the value of communication services beyond the bit pipes. 

Even their traditional core service of voice and text messaging, are going the way of dinosaurs. For example, WeChat in China has basically replaced all voice and message between users. The much heralded Rich Communication Service (RCS), has languished for years until recently revived by google merely as a competitive ledger to combat Apple’s iMessage.

Dean Bubley, a respected telecom analyst at Disruptive Analysis, provided us with this juicy quote “The telecom industry’s attempt to create a new messaging standard has been a 15-year odyssey of failure. Despite ample evidence from 2004 onwards that messaging was going to fragment into many use-case and behaviour-specific tools, it persisted trying to design the RCS camel by committee. Instead of individual MNOs – or small groups of them – experimenting and innovating, they instead leaned on GSMA to create a lower-than-lowest “interoperable” common denominator with zero focus on real user needs, or agility. Google’s attempt to turn RCS into first an iMessage clone, and then a B2C interaction platform, while simultaneously keeping the telcos onside, has just lengthened the requiem for this zombie technology.”

But the network operators, with their tremendous size and resources, are taking aggressive steps to establish themselves as the masters of both access and service realm. Acquisition of premium content owners and aggregations, such as Verizon buying Yahoo and AOL, AT&T buying TimeWarner, is one popular strategy. Reliance Jio of India, built its brand new 4G network to enable the largest mobile video streaming system in the entire world and put India in the forefront of what a future operator can aspire to become.

7. From perfection (guarantee, low speed, high cost) to simplicity (no guarantee, high speed, low cost)

This thesis is mostly true.

Even 4G/LTE and 5G are mostly about providing the highest throughput data pipe without much fuss. All the talk about QoS, class of services, IMS/RCS pales in comparison to the flourishing OTT services that cope with any connection they can get. 

And for most parts they work well, as long as there is reasonably abundant bandwidth and low latency. The only exception is the action game streaming, e.g. Google Stadia, where early adopters have reported disappointing 4K/HD performance with the existing infrastructure that connects user’s home to datacenters.

WiFi, as a simple best effort technology, is dominant indoors, for home, office, and the Starbucks in between. And WiFi has accounted for more than half of all Internet traffic. So good enough is actually enough.

Other big signals and macro trends

When we look back at the 15 years since our last essay, we did not foresee a few large trends.

• The battleground has shifted: telcos declined while the Internet giants (GAFA and BAT) rise.
• We did not foresee the Artificial Intelligence and Machine Learning train coming: the centralized platforms with tons of data are more powerful than ever, and that power is increasing every day with the data we all contribute. The barrier is less in algorithms but more in the collective data those companies have mined over the years. And the gap might be widening between the top Internet giants and the rest of the pack.
• Rage against the machines: Big corporations, powered by big data and AI/ML,  are in such a dominant position not only in deciding an individual's life but also nation state’s political system that humanity’s very existence felt threatened. The backlash against Facebook and other Internet giants, signaled the awakening of the “the useless class” defined in Yuval Noah Harari’s seminal book “ Sapiens”.
• Decentralization movement: on the heels of the 2007 financial crisis, Bitcoin was created to build a financial system for everyone that is not controlled by a few centralized entities whose trust has proven to be abused. Furthermore, in Chris Dixon’s proposed in “why decentralization matters”, that cryptonetwork might be the only potential alternative to those ever more powerful Internet giants. 

The road ahead

If we believe that data is the new currency and the higher intelligence should be shared by all humanity, then we can define the key themes for the future:

• Private ownership of data
• Secure communication and sharing of data
• Trusted and augmented intelligence

And along these new themes, we can envision the information system of the future:

• Pod: an independent unit of compute, storage and networking (Solid project by Tim Bernes-Lee) that is owned and operated by individuals. All data is fully owned by the individual, and can only be shared with permission and fine granularity.
• Massive peer to peer (p2p) network to connect the pods and transmit data among them securely and with confidence. It is infinitely scalable not only to the billions of humans, but also to the trillions of smart devices. It is decentralized and no single entity can control the flow of data.
• Federated and symbiotic human-machine learning. There are two points here: first of all the machine learning model is federated and only meta data are shared among the participating pods; secondly it is not human versus machine, but rather a collaborative system to make the best of both human and machine.
• And to support all of the above, we need a fair economic model for all the participants of the shared humanity. Ideally such a model will be defined in software and code, instead of over complex legal jargon that can be manipulated by the privileged professionals.

So what do all these mean for networking in the future? What should innovators explore towards NET 3.0? We believe these are some potential paths:

1. Design for the pod architecture, where compute, storage and networking are all in one. When building the future communication network, innovators will always bear in mind where the data will be generated, processed and distributed, and how the computing and storage of data will flow. 
2. Focus on software, data and algorithms. There are still plenty of radio technologies to improve and lots of new hardware to build, and the incumbent large vendors have spent lots of R&D resources on ASIC and specialized SoC to maximize performance while reducing power consumption and cost. Outside of the incumbent vendor space, innovators might want to focus their energy on software, data and algorithms which do not require 100M+ USD investment and 3-5 years of return on investment. but they will increasingly become fairly standard commodities. Take a look at what Open Compute Project has already done to the datacenter and what Telecom Infra Project is trying to achieve. Rakuten Wireless, the newcomer in Japanese mobile market, is building its entire 4G/5G solution on a fully virtualized cloud-native network architecture.
3. Strive for near 100% utilization. As google has achieved near 100% utilization on their WAN links between datacenters, innovators should look for solutions to achieve near 100% utilization for every Internet link. With or without 5G, we need to find a solution to unleash the 50-60% network capacity that is wasted at any second.
4. Adhere to the decentralization principle. The Internet is so successful mainly because it is a fundamentally decentralized system with peering and autonomous decisions. Otherwise, it just will not scale. So even when the innovators are striving for 100% utilization and the temptations of centralized smart algorithms are ever so strong, we should adhere to this winning principle of localized decisions that eventually form intelligent global outcome. A bit like Cellular Automata, or how your nerve cells work in your brain.
5. Leverage the ubiquitous base layer. With the advent of SpaceX Starlink, we might finally see a true global and ubiquitous base communication layer that covers everyone in the world. That constellation in the sky, combined with an equally powered meshed network on the ground, innovators can build their local or specialized coverage while still have global “roaming”. Even MVNOs like Google Fi can benefit from it.

15 years after NET 2.0, the expedition is still on. Welcome aboard!

Acknowledgement

The authors would like to thank Dean Bubley, Founder and Principal of Disruptive Analytics and Chris DePuy, Founder and Technology Analyst of 650 Group for providing insightful quotes. We would like to thank Claus Hetting of Wi-Fi Now for helpful discussions.