Thanks for sharing. That had my 'old man shouting at clouds' vibe, I like it.... On 19 December 2023 16:01:09 CET, Dave Taht via Nnagain wrote: >---------- Forwarded message --------- >From: Dean Bubley via LinkedIn >Date: Tue, Dec 19, 2023 at 5:05 AM >Subject: The 12 Lies of Telecoms Xmas >To: Dave Taht > > >During 2023, I've lost patience with some of the more outrageous statements… >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ >[image: LinkedIn] > >[image: Dave Taht] > > >Newsletter on LinkedIn >[image: Dean Bubley's Tech Musings] > >Dean Bubley's Tech Musings > >Analysis and arguments on wireless, telecoms, 5G & the wider futurism >landscape by @disruptivedean > >[image: Author image] > >Dean Bubley > >Tech Industry Analyst & Futurist @ DISRUPTIVE ANALYSIS | Influential >advisor & speaker with 25yrs+ in Telecoms Strategy, 5G / 6G / Wi-Fi, >Spectrum, Policy > >See what others are saying about this topic: Open on Linkedin > >[image: Newsletter cover image] > >The 12 Lies of Telecoms Xmas > > >During 2023, I've lost patience with some of the more outrageous statements >I've been seeing in my industry. There's a growing number of bald-faced >lies, and I've started calling them out publicly. > >Yes, I know that marketing and lobbying requires a measure of hype, >exaggeration and glass-half-full predictions and estimations. There are >areas where statistics or semantics can be skewed, but may contain a grain >of fact. > >Some problems lie with simplifications or lack of understanding made by >non-experts. While politicians are generalists and can sometimes be >excused, the wide use by the industry of *economists* to calculate supposed >costs or benefits of networks is often deeply flawed. > >Worse, I see many examples where purveyors of dodgy stats and >talking-points cite (and amplify) each other's nonsense. You know the sort >of thing - a laughable analysis of "GDP uplift" from a technology gets >blended with questionable forecast traffic volumes, then multiplied by >other flawed numbers to imply huge extra CAPEX or spectrum needs. > >Nobody stops to ask users, application developers or enterprises how >networks in the real world actually work or get deployed, or whether all >the clever AI, automation and virtualisation we're also hyping might reduce >the costs. > >Few people really look at headline numbers or arguments to see what is >cherrypicked, misrepresented, show correlation not causation, or just don't >pass a simple "smell test" for being realistic. > >Anyway... here's a dozen of the most egregious falsehoods in telecoms. >There's plenty more, but this seemed an appropriately festive number to >list. I'll add some links in the comments as well. Here's to a more honest >and truthful 2024! >Lie 1: Autonomous vehicles & robotic surgery need 5G > >Let's start with an easy one that long been debunked, although I still hear >it repeated by some in telecoms or policymaking circles, as well as some >self-appointed futurists who should know better. > >No, AVs don't need 5G, although they may use it for certain functions if >it's available and inexpensive. All the claims of TB of data created per >hour are irrelevant - 99.9% never leaves the vehicle. Most of the >processing & AI inferencing is done onboard. And cloud-driven 5G AVs would >need perfect coverage and capacity, which doesn't exist - good luck in a >tunnel, a car-park or a road in a remote region. Or on a highway in a lane >sandwiched between two trucks. > >5G is somewhat more important for remote-driven vehicles where a human >needs streaming video of the surroundings. > >As for robotic surgery - well for a start, most hospital operating theatres >are Faraday cages deep inside buildings, so they'd need dedicated wireless >systems. And the remote surgeon - and the robot - are likeliest to use >fixed broadband and a fibre LAN connection, not wireless. > >You remember the video of the banana undergoing "5G surgery"? That was a >proven lie, with a full analysis by The Verge. It didn't use 5G. >Lie 2: There is an "investment gap" for broadband networks > >2023 has been a year of preposterous lies from lobbyists, some telcos, >industry groups and even government agencies, trying to concoct arguments >for new regulations on cloud or content companies funding networks - or >perhaps more government subsidies. A common refrain is that there is a >"gap" in funding that needs to be filled by someone else. > >The problems are that the majority of investments required for fixed and >mobile broadband deployments are either: > > - > > Covered by market forces and normal commercial investment plans, helped > by existing government funding programmes. > - > > Exacerbated by arbitrary or poorly-defined "requirements" > - > > Not driven by growth in traffic volumes > - > > Based on old or inaccurate metrics and statistics of current network > status & investment schemes > >My friends at organisations such as Stratix consulting have done a good job >at debunking some Europe-specific claims. (Link in coments). Let's have >some more informed & truthful debate in 2024. > >Sidenote: some investments needed won't be done by the normal telcos >anyway. Indoor wireless owners, local FWA, fibre altnets, private networks, >satellite systems, neutral host & infraco / towerco CAPEX weren't even in >the discussion - if there *is* a new pot of cash, it's another lie to say >that legacy telcos are the only or best recipients. >Lie 3: 5G networks are (or will be) "ubiquitous" > >One of the common mistaken assumptions among policymakers and some others >in telecoms is that 5G networks - and their headline capabilities like >gigabit speeds and millisecond latencies - will be ubiquitous. > >Clearly, that's not the case, and was doubtful from the start of 5G. Rural >areas, small communities, indoor coverage & capacity, full connectivity on >railways & trains, industrial zones and others often lack public network >coverage, even from 4G. Even where there's a 5G logo lit up on a phone, >that may just be a thin layer of sub-1GHz spectrum or dynamic-shared 4G/5G. > >To be fair, this is perhaps more an example of ignorance rather than lying >in many cases, but it should have been clear to even the least tech-aware >person by now that the concept is itself proven untrue. > >It's also very consequential - application and device developers are the >ones that have been lied to, as much as end-users. Imagine spending money >on developing AR/VR game or headset, expecting low-latency 5G everywhere it >will be used. This type of hype has victims. > >It's also quite amusing to see the latest ITU IMT2030 (6G) recommendations >specify that ubiquity only refers to a defined coverage area, not >nationwide: "*The term 'ubiquitous' is related to the considered target >coverage area and is not intended to relate to an entire region or country*". >Well, that's one way to avoid lying in future - redefine the word itself to >mean something else that's easier an more convenient. >Lie 4: Mobile data traffic is growing "exponentially" > >Exponential is a mathematical term referring to an accelerating growth >rate. In almost all cases, mobile data traffic growth is now slowing. > >Most predictions suggest it settling down to maybe 20% year-on-year in >mature markets. And even that is mostly driven by MNO-driven decisions like >offering FWA fixed wireless services (which have 20x the consumption of >normal smartphone use), or inappropriate pushing of unlimited plans or >bundled / zero-rated video. > >I still see telco policy people deliberately overestimating traffic >forecasts, either to make arguments for more spectrum, or when lying about >so-called "large traffic generators" or risible "fair share" schemes. >Lie 5: Network traffic is "generated" by content / cloud companies > >I've previously written a full LinkedIn post calling out the term "large >traffic generator" as a clear lie. It's been one of the most-read that I've >published this year, which presumably implies *I* generated a lot of data >traffic personally. > >The reality is that Internet *users* generate traffic. They request movies, >play games, scroll timelines, download software updates, and read articles. >The fact I watch a GB of video traffic from a popular site rather than a >niche one is irrelevant - although the larger one is more likely to have >its own CDN. > >There's a small amount of "push" data such as autoplay follow-on videos, >but adverts are usually tolerated as an alternative to fees. It's not a big >deal - and in any case, network costs are linked to *peak* traffic levels, >not total volumes. >Lie 6: "Voice" is the same as "Telephony" > >There are 1000s of forms of voice communication. Phone calls are just one >highly-specified application, with specific behavioural, technical and >regulatory characteristics. Telephony is > >Push-to-talk, in-game chat, karaoke, audioconferences, voice assistants, >podcasts, audio captioning, Alexa "drop-ins" and many others are not >telephony. > >Generally, telcos don't do "voice" in general. They only do telephony, plus >a few extra bits like voicemail. A few have their own voice home assistants. > >(Historical sidenote: audio streaming using phone networks has been around >since the 1880s. If you're a telco complaining about "OTT" media content, >you're 150 years late). >Lie 7: All Wi-Fi use on smartphones is "offload" > >The term "offload" really involves Wi-Fi traffic that would otherwise have >gone over cellular networks, but an automated systems pushes to Wi-Fi >instead. It excludes data that the user (or the OS) deliberately selects >Wi-Fi for. > >If I watch a YouTube video on my sofa on my phone via Wi-Fi, that's not >offload. That's just me using my home broadband. > >Maybe only 5% of smartphone Wi-Fi data is genuine offload. And even that >could translate to a smaller amount over cellular, because both user and >app-developer choices often mean extra "elasticity" - higher resolution or >frame-rates, extra usage because it feels free/unmetered and so on. >Lie 8: Interoperability is always beneficial > >Interoperability for infrastructure can be very useful - we all benefit >from devices that work with wireless or fixed broadband networks, and are >tested and certified to be compatible with each other. > >But interoperability for applications and services is much more mixed. As >long as users can "multi-home" and have several different calling, >messaging, social or gaming platforms, it's not necessary to have >interoperability. > >Yes, there can be competition concerns, but that doesn't imply a need to >regulate for a lowest-common denominator set of features, with a wide array >of unintended consequences. Interop should be on-demand (especially if >customers explicitly ask for it), not assumed to be an ideal scenario and >mandated by regulation. > >Yes, I'm thinking of the EU DMA's stupid rules on messaging interop, but >also a more general argument about all communications apps & services, as >well as fields like AR/VR/metaverse. There's still a role for proprietary >network technologies and architectures in some cases too - openness is >great, but it's wrong to say it should be universal. >Lie 9: QoS or performance can be guaranteed "End to End" > >Almost every time you see the term "end to end", it's a lie, especially in >reference to network quality or performance. > >In reality, nobody controls *all* the components of a network or >application path, outside very niche and specifically-engineered >situations. At best, they can oversee and prioritise activity between two >arbitrary points in the middle. > >Think through a videoconference session. It involves at least two networks, >plus interconnections. Multiple devices, with multiple components like >processors and memory. A cloud platform and maybe a CDN and 3rd-party >elements. Maybe a last-metre connection via Wi-Fi or Bluetooth or a USB >cable. Maybe one user is deep in-building or in a moving vehicle. > >End-to-end QoS can be created if there's one network, IT or OT manager in >charge of a local system. It can be offered on one fibre connection and a >few boxes, or perhaps across one boundary. But truly end-to-end? Nope. >Lie 10: "Digital" tech is new, cool & all that matters > >Yes, I have a bee in my bonnet about the word "digital". I cringe at >phrases like "digital transformation", or "digital infrastructure". It's >like going back to the 1980s and hearing "information superhighway" or >"multimedia". > >We've had digital communications since Morse and the telegraph in the >1840s. Digital computers since the 1940s. Digital phone switches since the >1980s. The WWW since the 1990s. > >Digital is old. > >Don't get me wrong, it's still useful. (Obviously). But so is fire and the >wheel. > >What irks me is that much of the coolest stuff in tech today isn't digital >- it's quantum, or neuromorphic, or biological, materials-based, analogue >and many other domains of innovation that don't depend on 1's and 0's. > >The new & cool stuff is mostly *Post-Digital*. >Lie 11: Mobile enables huge CO2 savings > >There are many claims that use of mobile (or 5G) leads to huge savings in >carbon footprint - fewer flights because of videoconferences, wireless >traffic controls reducing congestion, connected solar panels and batteries, >and so on. > >Most of this is a massive exercise in double, or 10x-counting. If I do a >videoconference with a client, there's at least 2 access networks, >interconnects / long-haul fibre, two end-devices, two screens, two cameras, >lots of chips, a cloud platform, a video app or browsers, mics etc. > >Being generous, one of the access networks (probably fixed+WiFi rather than >mobile anyway) maybe accounts for 5% of the total savings. And only a >handful of trips are replaced with video, in any case - most video sessions >are incremental, or replace a lower-energy voice-only call. >Lie 12: "The data and KPIs prove XYZ" > >One of my big campaigns this year has been about the need for good metrics, >not easy metrics. I've written several posts about it, and an entire report >for my friends at STL Partners. > >In telecoms we get huge volumes of reported data - average network speeds, >5G coverage, spectrum prices in $/MHz-pop, homes passed, minutes of use, GB >of data traffic, numbers of messages sent, and myriads more. > >The problem is that many of these data points are used because they are >easy to collect, easy to regulate & have a lot of historical comparisons. >But they're often useless, or sometimes worse-than-useless. > >They get cherrypicked, used out of context, used to "prove" the need for >policies that they don't actually support, and don't represent the reality >on the ground. > >Take mobile coverage, usually cited as a % of residential population that >can get a signal outdoors. Yet most will agree that many 5G uses are for >non-residential applications and locations (eg farms, ports, industrial >zones), or indoors. And coverage doesn't mean performance or "full 5G" >capabilities. > >Then there are all the many tiers of lies about network data traffic, used >to "prove" the costs of network CAPEX, energy use, need for taxing >cloud/content companies, arguing against net neutrality and so on. >Aggregate traffic is largely irrelevant - in reality costs are driven by >initial build / coverage area (even at zero traffic) and sometimes peak >rates. Energy usage isn't linked to traffic mostly, either. > >We all know the phrase "lies, damned lies, and statistics". So let's try to >collect and use better stats in future. >Conclusion > >Yes, this is a bit of a rant. And yes, a lot of people will object to me >calling them liars. But let's be clear, most of those I'm referring to have >jobs to do, that sometimes need "messaging" to be massaged. > >And I haven't even got onto the dodgy term "OTT", mentioned RCS (no it's >not SMS2.0), laughed at cryptocurrency-powered "DePIN", called out >exaggerations about the role of satellite direct-to-device, or ridiculed >the criticism thrown at shared spectrum, CBRS or private networks. > >So if you recognise yourself in this, tell the real telco Pinocchios - the >lawyers, lobbyists, marketeers and headline-writers - that you're not >willing to peddle untruths or half-truths in future. Don't let them put >your nose out of joint, or make it grow longer. > >Join the conversation > > >Know someone who might be interested in this newsletter? Share it with >them. >[image: Share this series on LinkedIn] > >[image: >Share this series on Facebook] > >[image: >Share this series on Twitter] > >This email was intended for Dave Taht (@dtaht:matrix.org - Truly speeding >up the Net, one smart ISP at a time) >Learn why we included this. > >You are receiving LinkedIn notification emails. Others can see that you are >a subscriber. >Unsubscribe > > · Help > >[image: LinkedIn] > >© 2023 LinkedIn Corporation, 1‌000 West Maude Avenue, Sunnyvale, CA >94085. LinkedIn >and the LinkedIn logo are registered trademarks of LinkedIn. > > >-- >40 years of net history, a couple songs: >https://www.youtube.com/watch?v=D9RGX6QFm5E >Dave Täht CSO, LibreQos -- Sent from my Android device with K-9 Mail. Please excuse my brevity.