Rakuten Symphony exec: “5G is a failure; breaking the bank; to the extent 6G may not be affordable”
In a Linkedin post, Rakuten Symphony CMO Geoff Hollingworth says 5G has been a failure, that it will never achieve ubiquitous coverage and it’s time for the mobile industry to invest in customers rather than networks:
5G is a failure. We build technology to deliver a promise, the promise made was commercial. If that promise is not delivered then it has failed. If that promise is delivered but the cost of delivery is prohibitively expensive versus the return, then that technology is a failure. Currently the promise has not been delivered and there is no line of sight to delivery.
I do not believe the market is going to deliver on this promise further down the road, the networks do not work this way. 5G will never have ubiquitous coverage, and this is getting worse. We can get closer to ubiquitous coverage as a network of networks but then all the complexity embedded in 5G for advanced management of the network is a cost. There is never only one “G” in a market, there tend to be at least 3 at any moment. Each G takes time to retire, from a customer, device, and ROI point of view. APIs cannot get deployed universally and by the the time networks are universal the next G is starting to be rolled out. We are still waiting for “real 5G” but even when it shows what will it translate to commercially? Will it arrive before something labeled “6G” is starting to roll out?
“The current cost for 5G is breaking the bank, to the extent where 6G might not be affordable. None of this is good for an industry that is supposedly powering the global GDP and defining the future state of all countries.”
Of course, we agree as we’ve been pounding the table since the 3GPP Release 15 pre-standard version of 5GNR was introduced in 2018. The “real 5G” must include standardized 5G SA core network, URLLC that meets ITU-R M.2410 performance requirements and harmonized frequencies.
“We are already repeating the narrative into 6G with the same concept of ubiquitous coverage mindset. We are not segmenting for actual coverage and actual market reality when comparing to cost, time, and need,” Hollingworth warns. “We must embrace that there are other better ways to solve for coverage, depending [on] the use case and the coverage type required. We have less of a coverage problem and more of a seamless access problem, as one moves from indoor to outdoor city to outdoor suburb to rural.”
…………………………………………………………………………………………………………………………………………………………………………………….
Future Network Trends -AI native and cloud native operations:
To see future network needs, there is a need to understand future AI software design patterns. We now see these appearing and they are different from what we have seen before.
- They depend on large data streams where the data and interpretation has time sensitivity.
- Compute is distributing to where the data is rather than bringing the data to where the compute is.
- This forces distribution of models and software to where the data and compute is, and overhead must be kept to a minimum and automation must be maximized to be cost effective.
- There is need for rapid iteration and continuous fine tuning of the models and algorithms as more data is analyzed and performance improves.
The first applications with these design requirements are the 5G++ network architected functions. The test of the industry is whether the software can be adopted with true AI native, cloud native operations or whether they will be deployed traditionally. If we succeed in deploying a true hyperscale operation for our own software we can take our tooling and knowledge, and expand it to support any application and service.
Now there are many networks that devices can connect to, and 80% of the time traffic does not travel through cellular networks at all. In markets with high fiber penetration Wi-Fi connectivity and latency has higher performance than cellular, especially indoors, where higher frequency spectrum no longer travels inside buildings, especially those with reflective sustainability materials.
We previously solved as if cellular was the only way to solve all the problems. We must embrace that there are other better ways to solve for coverage, depending the use case and the coverage type required. We have less of a coverage problem and more of a seamless access problem, as one moves from indoor to outdoor city to outdoor suburb to rural.
References:
https://www.linkedin.com/pulse/cost-delusion-promise-reality-geoff-hollingworth-vt4xe/
https://www.itu.int/pub/R-REP-M.2410
https://www.lightreading.com/6g/jumping-off-the-g-train
5G has not produced any material growth in sales for its network operators:
Evidence of a 5G identity crisis can be found in the UK Competition and Markets Authority’s (CMA) ruminations on a proposed merger between Vodafone and Three. According to third-party feedback on aspects of the deal, 5G “standalone,” the version with the boost of a new core network, “remains a nascent technology that has yet to develop widespread use cases,” it said in a detailed paper. Consumers might not be able to tell the difference between that and 4G, wrote the CMA.
There are signs of telco apathy toward 6G after the capital-intensive and underwhelming experience of 5G. Last year, the Next Generation Mobile Networks Alliance (NGMN), a club of big operators, published a position paper that said “6G must not inherently trigger a hardware refresh of 5G RAN [radio access network] infrastructure.” Where possible, said the NGMN, operators must be able to introduce the new technology via “software-based feature upgrades of existing network elements.”
Apparently sympathetic to that stance, Jan Ellsberger, elected director-general of the European Telecommunications Standards Institute (ETSI) agrees 6G is likelier to be an evolution of 5G than a revolutionary change. “Yes, I think that is the sentiment right now. That is most likely how 6G would be developed – improvements to 5G and securing that we deliver on the opportunities that are part of the 5G discussions already.” From a use case perspective, 6G, like 5G, might have more relevance for certain industries than it does for individual consumers, he thinks.
But on the technology side, he does not anticipate a radical departure from systems such as orthogonal frequency division multiplexing (OFDM), integral to the air interface used in 4G and 5G. “My prediction is that we will continue to carry on with the same type of base technologies but enhance them with new advanced modulation methods, maybe coding schemes, waveform design, also maybe even more advanced MIMO [multiple input, multiple output] configurations,” he said.
If the NGMN’s wishes are realized and 6G – driven by software upgrades rather than hardware replacement – necessitates less capital expenditure than its predecessors did, it could further squeeze revenues for ailing manufacturers like Ericsson and Nokia, the big European contributors to the Gs. But it would not alter the role of ETSI and other standards bodies as the effective guarantors of interoperability, says Ellsberger.
He also brushes off talk of Europe falling behind in standards development. “The tech players we have in the industry have a global market and are not serving only the European operators. They are serving globally,” he said. “From that point of view, in technology innovation and standardization, we are still the leading region in the world in this space.” If there is a technology lag, it is in the deployment of 5G and caused by the fragmented nature of the European industry, as far as Ellsberger is concerned. It is a view that senior executives at Ericson and Nokia would share.
https://www.lightreading.com/6g/don-t-expect-a-revolution-with-6g-says-etsi-boss
While 5G (and to a lesser extent 4G) already is undergoing some convergence with non-terrestrial networks (NTNs), Ellsberger said we can expect this to be taken a step further in 6G.
“There will be a tight integration and by that you can use satellite as a complement to have a wider coverage or better coverage than what we can get with pure land-based communication.” he said. “It is important you have to emphasize that 6G is an evolution of 5G and not that disruptive technology. So, I mean whatever is done when it comes to NTN or satellite integration of course have to respect that ambition.”
Thinking back at innovation with 5G and the killer app, we are reminded of how much of the hype and buzz at 5G’s launch was driven by the marketing machinery. And despite the hefty investments in spectrum, in 5G infrastructure, and ongoing investments in 5G-Advanced with release 18, the sales and revenue generating vehicles have left much to desire for.
https://www.telecoms.com/5g-6g/etsi-boss-says-6g-an-evolution-of-5g
Splitting 5G into two parts
1:access network (radio)
more bandwidth and greater number of terminals, works good
2:network core (MPLS/IP), the problem
3 In my country, two remote operations were carried out using 5G, this never leaves the laboratory environment and cannot be transferred to production. Why ?
Because MPLS/IP switching is not capable of maintaining synchronous video and voice in both directions, a digital circuit (with ip ) was used during operations.
The number of digital circuits in a backbone is very small the service cannot be sold
4 A B
Video –> <– Voice
Synchronous in both directions. Anyone who knows a little about computers knows, that you need a kernel with semaphores to do this. Ruters cannot have a kernel because they need low latency.
seee:
Average packet processing time in VPP, (Cisco algorithm — Average latency: <23 usec)
https://openstack.fr/r/meetup23/FDio-the-fast-dataplane-Jerome-Tollet.pdf
Latency is measured in milliseconds, with latency time in micro, the latency of circuit switching is close to the latency of packet switching.
But digital circuit switching allows a FIFO between A and B that is Full Duplex, and allows synchronizing voice and video
5 When an old technology that is 30 years old plus a new one, Wi-Fi, do what 5G promised and 5G doesn't, there is a big problem.
6 If they hadn't turned off ISDN it would have been possible to command robots anywhere in Europe : https://www.youtube.com/channel/UCAmLlT2YKjnTc8Lbkuw6jcg
7 All the advantages of SIP over ISDN are true but they are only for voice, they are not true in video and there is nothing that prevents SIP from running over ISDN
8 In conclusion, 5G works well, the problem is the core of the network, when they put SIP/Video on broadband digital circuit switching,
chosen by the client's SIP signaling, we will have quality video conferences, etc.
9 And keep MPLS/IP core for Voip and Internet