Showing posts sorted by date for query bandwidth consumption. Sort by relevance Show all posts
Showing posts sorted by date for query bandwidth consumption. Sort by relevance Show all posts

Wednesday, March 26, 2025

Asking the Wrong Question about 5G

The claim  that "5G has failed” is in some ways an odd one. On one hand, critics tend to cite the unfulfilled promises of exciting new use cases. On the other hand, critics tend not to focus on the lower latency, faster speeds or energy efficiency that each successive network also is founded upon. 


But that might be the main point: each successive mobile generation has been successful and necessary precisely for the reasons that consumer home broadband experiences have been based on ever-increasing bandwidth, capacity and access speeds. 


So alter the question just a bit to understand the real impact. Do you ever really hear observers arguing that mobility services (mobile phone service) actually have failed? One does not hear such claims because mobile service clearly has been a raging global success. 


Some 71 percent of humans presently use a mobile phone, according to the GSMA.  


source: World Economic Forum 


So “mobility” has clearly succeeded, even if some feel particular mobile platforms have not. To be sure, proponents have touted the creation of platforms for futuristic use cases (the network will support them), not the extent of usage. Some examples can always be cited, though often not mass market adoption. 


To be sure,  every mobile generation since 3G has made such claims. And we might advance some very-practical reasons for the claims. Each mobile generation requires the allocation of additional spectrum from governments, which have to be convinced to do so.


Pointing out the new potential applications; the contribution to economic growth; educational advantages and so forth are part of the effort to secure the new spectrum. 


Also, infrastructure suppliers have a vested interest in enticing operators to create whole new networks precisely because it might be possible to create new revenue streams, or provide


Still, each successive mobile platform has promised, and delivered, latency improvements of about 10 times over the preceding generation, as well as potential bandwidth (internet access speeds) of 10 times more, and typically also energy consumption efficiencies as well. 


The practical improvements always vary from laboratory tests, though. The actual behavior of all radio waves in real-world environments is an issue. So are the realities of impediments to signal propagation (walls, trees, other obstacles) and signal interference.


Cell geometry also matters. Higher bandwidth is possible when smaller cells are used. 


Higher bandwidth is possible when channel sizes are increased (as when channels are bonded together to create a single wider channel from two or more narrower channels). 


And real-world “customer-experienced speeds” also are dependent on which actual frequencies are used widely by each mobile generation. Lower frequencies propagate better, but higher frequencies support higher speeds, all other things being equal. 


Still, the point is that observers never question the “success” of the mobile phone and mobile networks, only the “failure” of futuristic apps to emerge. 


That is not the point. The primary and essential value of each successive mobile platform comes from network performance (lower latency, higher bandwidth) and not the possible new apps, which cannot be created by mobile operators in any case, anymore than internet service providers having created Facebook. Google, Amazon, YouTube or Uber. 


Mobile operators can only create the physical infrastructure third parties can use to create new use cases. And that has been accomplished. But then innovation leading to new apps rests in the hands of entrepreneurs and investors.  


That’s the whole implication of “permissionless innovation” the internet is based upon: innovators do not have to own networks to build apps that use the networks. The entities that own the access or transport networks do not necessarily or primarily create and own the apps. 


Oddly, the reverse tends to be the case: highly-successful consumer app providers find they can vertically integrate into core network transport as a means of lowering their costs. That is why most of the world’s long distance networks (subsea, especially) are built and owned by a relative handful of big app providers such as Alphabet (Google) and Meta. 


It is fair to note that few of the futuristic apps touted for 3G, 4G or 5G networks have become mass market realities. On the other hand, lots of highly-useful apps not envisioned for any of those networks have emerged.


Net

Predicted "Futuristic" Use Cases

Unexpected "Everyday" App Developments

3G

Video conferencing, mobile TV, advanced multimedia

Mobile social media (early stages), basic GPS navigation, early app stores

4G

Immersive VR/AR, high-definition mobile gaming, remote surgery

Ride-sharing apps (Uber, Lyft), widespread video streaming (YouTube, Netflix), robust social media (Instagram, TikTok), advanced turn-by-turn navigation (Google Maps)

5G

Holographic communication, tactile internet, massive IoT deployments

Enhanced real-time location based services, very high definition mobile video streaming, cloud gaming, very reliable real time social media interactions. Increased use of live streaming services, and the further enhancement of cloud based applications.


All of which suggests we are very bad at predicting the future; innovations often emerge unexpectedly and only when users see the value. 


Consider only the industrial, commercial, medical and other applications generally centered around the use of sensors and mobile networks as the connectivity mechanism. Most have not taken off in a significant way, even if there are some instances of viable and routine deployment. 


Generation

Touted Possible New Applications

3G

- Telematics for automotive industry5


- Smart home devices (thermostats, security cameras)1


- Traffic light systems1


- Vending machines with remote monitoring1


- GPS trackers for livestock1


- Wearable devices and e-readers1


- Medical alert devices1


- Remote weather stations1

4G

- Enhanced mobile broadband for video streaming and gaming6


- Smart home applications2


- Internet of Things (IoT) connectivity2


- Remote monitoring systems2


- Vehicle communications (real-time road information, navigation)2


- VoIP calls and video conferencing6


- Mobile payments6

5G

- Telesurgery and remote medical procedures4


- Fully autonomous vehicles4


- Advanced connected homes4


- Portable Virtual Reality (VR) experiences4


- Smart city infrastructure4


- Ultra-reliable low latency communication (URLLC)3


- Massive Machine Type Communication (mMTC)3


- Industrial automation and robotics8


- Remote patient monitoring in healthcare7


- Large-scale IoT deployments in agriculture, utilities, and logistics


For the most part, the futuristic appl;ications have not developed as expected, and when they do take hold, it often is in the subsequent generation.


Many expected 3G to produce mass market usage of videoconferencing. That did happen, but only in the 4G era, with social media and other multimedia messaging apps, for example. That is a fairly common pattern: we overestimate routine adoption by at least a decade. 


Use Case Prediction

Actual Adoption (at least early stage)

Delayed Applications Likely Emerging in Later Generations

3G Expectations

(Medical devices, telematics, mobile TV)1

4G Realizations (IoT connectivity, smart meters, vehicle telematics)2

4G Concepts for 5G Era

- Advanced industrial automation3

- Mobile medical monitoring systems3

- Smart grid controls3

- HD public safety cameras3

4G Expectations

(Massive IoT, Industry 4.0)2

5G Realizations (Network slicing, enhanced mobile broadband)4

5G Concepts for 6G Era

- Holographic communications5

- Autonomous vehicle networks57

- Network-as-sensor technology5

- Microsecond-latency telesurgery7

5G Expectations

(URLLC, mMTC)34

6G Projections

- 1,000x faster latency than 5G7

- AI-optimized networks5

- Energy-efficient massive IoT6

6G Horizon

- Real-time digital twins5

- Military-grade AR simulations5

- Advanced environmental sensing5

- 8K holographic streaming


The point is that mobile services and smartphone services have proven wildly successful. In fact, nobody doubts that. What often gets criticized are the many futuristic apps that could be developed with each next-generation mobile network.


That misses the point. As fixed network home broadband has to continually extend internet access speeds and bandwidth, so too do mobile networks. The bottom line is that each successive mobile generation succeeds to the extent it does so.


Tuesday, February 4, 2025

Jevons Effect Applies to DeepSeek, Not Nvidia

Many financial analysts and investors are concerned that lower-cost language models such as DeepSeek will reduce demand for high-end graphics processor units, servers and high-performance computing services. The drop in Nvidia market value by about $600 billion after the DeepSeek unveiling of its latest model, which, it is claimed, was developed at a fraction of the cost of other leading generative AI models, attests to that fear. 


Others do not believe a dip in need for high-end Nvidia servers will happen. The Jevons Paradox or Jevons Effect has been cited in this regard. The Jevons Effect, also known as the rebound effect, is the proposition that increases in efficiency in resource use can lead to an increase in the consumption of that resource, rather than a decrease.


So the reasoning is that cheaper AI models and cheaper inferences will lead to much more AI use,and therefore continued demand for Nvidia servers, even high-end machines. Some might point out that the claimed DeepSeek cost advantages should properly mean the Jevons Effect happens to DeepSeek and other models, not the servers running those models. 


So the reduction in demand for Nvidia servers would be indirect. That has not deterred many observers, who believe significantly-lower-cost model training and inference operations are possible (with or without DeepSeek contributions). We already have seen such effects.

source: a16z

 

In principle, some argue, such trends might mean model operators could reduce the need for high-end graphics processing units and other acceleration chips. That remains to be seen. 


Broadly speaking, the history of computing suggests the “lower costs lead to wider usage; more usage; more demand for servers” argument has legs. Computing resource demand always has increased as the cost drops. 


But there also is an argument that lower-cost models and inferences will increase demand for all those products. Paradoxically, lower training and inference costs will stimulate more demand, possibly lead to new use cases 


Also, keep in mind that generative AI is only one form of AI. Perhaps SeepSeek points the way to lower processing costs for GenAI. That does not necessarily mean improvements for general AI or even machine learning. 

One might argue it remains necessary to use high-end processors and accelerators for use cases moving towards artificial general intelligence. 


The Jevons Effect, or Jevons Paradox suggests that an increase in efficiency in resource use can lead to an increase in the consumption of that resource, rather than a decrease, as some might expect, over the long term. 


Industry/Resource

Efficiency Improvement

Observed Increase in Consumption

Coal (Historical Example)

More efficient steam engines in the 19th century

Increased coal consumption as steam engines became more widely used

Oil & Gas

Fuel-efficient car engines

More people drive longer distances, leading to sustained or increased fuel demand

Electricity

Energy-efficient lighting (LEDs, CFLs)

More lighting is used due to lower energy costs, offsetting savings

Computing Power

More efficient processors (Moore’s Law)

Increased use of computing applications and data-intensive services

Water Usage

Efficient irrigation systems

More land is cultivated, increasing total water use

Internet Bandwidth

Faster and cheaper data transmission

More streaming, gaming, and cloud computing, increasing total bandwidth consumption

Paper Consumption

Digitalization and efficient printing

More documents are printed due to lower per-page costs

Ride-Sharing & Public Transit

Efficient, affordable transport services (Uber, electric buses)

More trips taken, sometimes replacing walking or biking

Air Travel

Fuel-efficient aircraft

Lower costs lead to increased air travel demand

Food Production

High-yield farming techniques

More food is produced and consumed, increasing overall agricultural resource use


As applied to more-affordable generative artificial intelligence solutions, a drastic decrease in cost might-also lead to a decrease in buying of high-end AI servers, for example, even if model training can be done on lower-capability servers. (at least in some cases). 


The argument is that efficiency doesn’t reduce demand—it increases it. As AI becomes cheaper and more accessible, more businesses, startups, and individuals will adopt it and new higher-end use cases will emerge, driving more higher-end GPU sales. 


Perhaps the bigger immediate concern is that many contestants have essentially overpaid for their infrastructure platforms. 


That becomes a business issue to the extent that, in competitive markets, the lower-cost producers tend to win. 


On the other hand, the Jevons Effect works when the price of the inputs does not change. If the price of high-end Nvidia servers does drop, then supply and demand principles--and not the Jevons Effect--will operate. And lower prices for high-end Nvidia GPUs then sustains demand. 


The Jevons Effect suggests that improved product efficiency leads to greater overall resource consumption rather than reductions. That might apply to use of AI models, AI as a service or power consumption. 


But  many have speculated that AI models such as DeepSeerk would lessen the need for high-end graphics processor units, for example. That might hold only if the prices for such servers remains the same. 


And the general rules of computing economics suggest lower prices with scale and time. So the ultimate impact of DeepSeek and other possible model contenders on demand for high-end Nvidia servers might be more neutral than some fear.  


And, in any cases, one might argue any effect on high-end server demand might affect GenAI models more than broader and more complicated artificial general intelligence operations.


What Will DoJ Require of Google?

Google is now in the “remedies” phase of an antitrust lawsuit filed by the U.S. Department of Justice, as Google has been found to be a mono...