A16Z Partner: 5G is coming. What can we do?

 A16Z Partner: 5G is coming. What can we do?

However, Benedict Evans, a well-known investment partner of A16Z, puts forward a different point of view in this paper: the increase of network speed does not necessarily lead to the popularization of VR and auto-driving applications, and AR may benefit greatly from 5G network. Just as the 3G era has not brought about a boom in paid video telephony, the real phenomenal App of the 5G era will include services that people expect to achieve, but it should go beyond that.

The original text is from Ben EvansBlog, author Benedict Evans

To be honest, unless you work in a professional field with a high degree of relevance to 5G, you shouldnt spend too much time thinking about these issues.

In early 2000, European mobile operators spent 110 billion euros to buy 3G licenses under the influence of Internet bubble, mobile foam and broadband bubble. Now, almost 20 years later, when I came back from the International Consumer Electronics Show (CES), 5G became a hot topic.

Many friends in big companies tell me that the question of what is 5G has spread to every corner of the headquarters of many companies, almost as common as asking what is machine learning.

There are many different ways to answer this question.

Ill talk about network efficiency, reconfiguration, vendors, Huawei, chipsets, and perhaps network function virtualization (NFV). Its a concept of network architecture that uses virtualization technology to divide the functions of the network node hierarchy into several functional blocks, which are implemented in a software way and are no longer limited to the hardware architecture.

But Im no longer an analyst in telecommunications. Now I work in Silicon Valley. So, from the perspective of Silicon Valley, I think it can be analyzed from four aspects.

What practical changes can 5G bring?

Without (too much) talking about technical details, lets see what people can get from 5G. The answer is a wider pipeline.

Like every previous update, 5G allows mobile operators to expand in a cheaper and easier way, thus continuing to adapt to increasing network usage.

5G will be deployed on the spectrum of existing cellular radios, but it will also allow operators to use higher radio spectrum resources (more than 20 gigahertz, also known as millimeter wave or mmWave), which have never been used for mobile services before (and many short-range base stations will need to be installed in the process). ?

Because of the deployment of this new spectrum, under good conditions, mobile 5G may travel at speeds of well over 100 Mbps and possibly hundreds of Mbps (in theory, mobile network speeds may exceed 1 Gbps, but are unlikely to be achieved in the real world).

5G is expected to achieve shorter network latency than 4G - in the real world, it may be 20-30 milliseconds, while LTE or 4G will take 50-60 milliseconds. It is not clear how obvious this improvement is for users.

Some people, such as Verizon, a telecom operator, argue that you can also use high-frequency spectrum 5G in home broadband services (meaning you need an antenna outside your home or on your window), with a network speed of up to gigabytes per second. However, many people questioned the proposal economically and technically.

Of course, the latest version of the Data Over Cable Service Interface Specifications (DOCSIS) means that the cable Internet can provide the same service - about a third of Americans can reach gigabit/second speeds.

Therefore, the network delay time of mobile devices will be shorter, and the mobile network pipeline will be wider and wider. In some places, fixed broadband will face more competition. ??????????????


What does a wider transmission pipeline mean? ?

Digital Subscriber Line (DSL) and the first deployed 3G provide us with hundreds of Kbits/seconds of transmission speed; upgrade to 3G (3.5G) and 4G makes the transmission speed reach tens of Mbits/seconds (and shorter delay time); upgrade to DSL and DOCSIS makes the fixed home broadband speed reach tens or hundreds of Mbits/seconds.

Each surge in transmission speed results in two outcomes. First of all, what we are already doing is getting smoother, easier and wider. At the same time, web pages can brush more images and become more dynamic.

Secondly, new things are possible. You cant use Flickr or Google Maps via dial-up, or access Netflix (at least inconveniently) via broadband in 2003.

In previous networking environments, only if you assumed that all users (without home WiFi) could reach a speed of tens of Mbits/seconds, can you effectively use the chat software Snapchat. This in turn means that a strong overall network at the same time can not only provide a considerable network speed for a single person, but also for many people at the same time, and can provide people with network infrastructure to support this speed at a reasonable price. If you show Snapchat to a mobile network executive at the beginning of the 21st century, he will be shocked because the early 3G simply cant bear the load.

Similarly, 5G speed and faster home broadband will mean that existing applications will become richer and new applications will emerge - New Flickr, YouTube or Snapchat. The creativity of entrepreneurs and platforms, as well as the choices of consumers, will determine the final outcome, although we do not yet know what the specific outcome will be.

This is the greatest innovation that can occur without permission in a decentralized Internet -- telecom companies dont need to decide in advance on use cases (use cases are descriptions of how systems respond to external requests in software or systems engineering, a technology that captures requirements through user use scenarios), just as Intel doesnt need to be a faster central processing unit (CPU). Cessingunit (CPU) provides the same use case.


AR, VR and automobiles

Although we dont know what kind of use cases will appear, there are some words in 5G related topics: AR, VR and autopilot. Its worth taking some time to discuss these areas.

I think VR is essentially an indoor product - you dont use it when you walk down the street, and you dont take it out for 20 seconds when you wait for a bus. This means that your home broadband type (DSL, optical fiber, cable, or maybe 5G) and connection mode (mostly via WiFi) constitute your VR connection mode.

Five gigabytes means two different possibilities: Fixed five gigabytes (very limited coverage) connected to an antenna outdoors at gigabit/second speeds and connected to a headset via WiFi, or a cellular mobile modem in the device, where the network speed will be closer to todays four-gigabyte LTE (again, 20 Ghz signals cannot penetrate the wall). But the network latency will be shorter than 4G.

Is 5G better than the existing connection? Will you notice the change in 5G usage? Is the breakthrough for VR waiting 5G? How many applications use 5G?

5G appears to be more interesting for AR. First of all, it should be clarified that todays AR is used to describe three different things:

- When you wave your cell phone at an object, you see the object on the screen.

- Wearable head-up display (Google glasses), but can not perceive the surrounding environment after wearing;

- A transparent, immersive, full-3D color display with an induction kit can be used to map the rooms around users and identify the corresponding people and objects. Companies, including Magic Leap, an AR startup invested in A16Z, are working hard to achieve this goal - and commodities will take years to reach the mass market.

The third description is the most interesting one for me. If you wear a pair of glasses, through which you can see the real world, and at the same time it can give feedback to the virtual world according to your operation, it will be very practical.

Just as the 3G network puts the Internet in peoples pockets, the 5G network will prompt all kinds of unpredictable new things to happen (imagine what it would be like to sell Snapchat if our only Internet experience was to connect on a PC by dialing up). Although AR can operate in a 4G network, the continuous low power consumption, high network speed and short delay of 5G will make the connection more effective.

On the other hand, I also heard about the application of 5G on autopilot cars. Im not sure if this proposal is feasible. Autopilot will surely require a lot of data, download the real map (real, very high-definition 3D model, show the street where their car is), and update the maps with the data of their own sensors. The car will download and update its driving system and upload more data about real human driving.

However, these download updates hardly need to be done in real time - they can be done nightly or even weekly. Mainstream autopilot cars do not require continuous connection, let alone 5G connection. Autopilot must work without a cellular service network.

After further observation, we can find that in a world with many self-driving cars, people are generally interested in Vehicle-to-Vehicle (V2V) technology. Using 5G or other wireless technology, automatic driving can achieve a higher level through V2V technology, such as passing through intersections without deceleration.

However, most of these use cases can only be achieved in a world without human driving -- obviously, this goal is difficult to achieve. When autopilot fails, or when you switch from manual to autopilot but autopilot doesnt work, we need remote control more urgently.

This is particularly closely related to truck transport: in the United States, about 90% of long-distance trucks travel on highways, which are easier to drive by themselves than on suburban streets.

Many companies are therefore exploring a model in which trucks drive automatically on the highway and humans take over the task of driving off the highway - either by entering the vehicle or by remotely controlling it. This would be a good 5G use case.


A cooler feature of 5G is that it allows you to partition dedicated capacity for specific use cases -- so-called network slicing. Nowadays (the real situation is much more complicated), although network operators try to regulate traffic, all traffic in a cell uses basically the same capacity. 5G allows us to create dedicated private capacity with specific characteristics in radio networks.

So you can sell truck operatorsdedicated capacity within two miles between a specific highway exit and a specific warehouse. Alternatively, you can provide a lower-priced but broader bandwidth for an Internet of Things (IoT) operator (or security company). ??????

Therefore, in theory, you can customize any combination of capacity consisting of data transmission speed, coverage, quality, delay time and reliability, and even incorporate power consumption into the reference elements (which may be interesting to discuss in the field of the Internet of Things).

There may be resellers who aggregate and implement these services on behalf of mobile network operators (MNO). This seems interesting - it seems to be a change in business and vertical services rather than consumer products.

We have seen a trend in this area: large business organizations use private 4G networks rather than WiFi or Ethernet. The same problem will also exist in 5G applications.


What are 5G phenomenal applications?

Facts have proved that putting the Internet in your pocket is phenomenal in itself. With the passage of time, video calls have become a part of it, but it is not a separate telecommunications service charged by telecommunications companies. Similarly, 5G phenomenological applications may be faster 4G.

Over time, this will mean the emergence of new Snapchat and YouTube, which will use new ways to fill pipelines that are not currently available, as well as new entrepreneurs. This may not be a revolution, or it means that the revolution that has been going on since 1995 will enter the next 10 years or last longer until 6G appears. ?

Source: Off-stack Responsible Editor: Sun Yunuo_NBJS9209