Winning the Race to Nationwide 5G Connectivity Matters in the Long Run

[cageymaru]

China and the USA have staked their bets on who is going to be more technologically advanced society in the future. China created their "Made in China 2025" initiative where they seek to upgrade Chinese industry and challenge American industry and their technological leadership. Deloittie Consulting has released a report on 5G that explains why it is so important for a country to be first to expose their citizens to new technology such as 5G.

It is speculated that an increased macroeconomic impact will be felt for those that early adopt 5G technology as it will create a "data-network effect"; which creates exclusive new markets within the early adopter country and thus leadership within that market. America has heard the alarm bells and are now rolling out coverage for the new wireless network in some cities, but industry leaders caution that we need more than 25X the current number of 5G cell towers in operation for proper 5G connectivity. Also China has already spent $24 billion on infrastructure since 2015. Thanks scojer for the news.


But beyond consumer applications, the tech could support connected infrastructure in cities, including driverless cars. In 2035, 5G is expected to enable $12.3 trillion of global economic output, IHS Markit estimated in a recent report. China has built 350,000 new cell sites, while the U.S. has built fewer than 30,000 in the same time-frame. "China and other countries may be creating a 5G tsunami, making it near impossible to catch up," Deloitte said in its report.

 

[vegeta535]

Yet still with 3gb data caps in 2035.

 

[pcgeekesq]

I disagree. Low orbit satellite communication constellations are going to make 5G irrelevant.
They are cheaper and cost less to keep running than the equivalent terrestrial capacity.

We'll see cell towers being taken down en masse before the end of the 2020's, with only regulatory inertia (i.e., preserving the ability of old devices to make phone calls) as a barrier to it.

 

[IdiotInCharge]

I disagree. Low orbit satellite communication constellations are going to make 5G irrelevant.
They are cheaper and cost less to keep running than the equivalent terrestrial capacity.

We'll see cell towers being taken down en masse before the end of the 2020's, with only regulatory inertia (i.e., preserving the ability of old devices to make phone calls) as a barrier to it.

I see them working side-by-side. Satcom will still be affected by interference, and connectivity will still be better between fixed ground stations than mobile radios. Meaning that while mobile to satcom will absolutely be viable and likely see widespread use, utilizing local terrestrial groundstations will still provide better overall service.

[and if you've tried to get a GPS lock on a device that lacks any 'helper' radios i.e. AGPS inside a concrete jungle or other interference-ridden environment, you understand why relying solely on low-earth-orbit infrastructure isn't a great idea]


In response to the OP: the competition here isn't likely to produce a 'winner'; in reality, we should expect both major competitors (US and China) to win together, even if they only cooperate as necessary rather than by plan. Further, this competition is going to drive up production and decrease unit cost, meaning that these technologies and the lessons learned in their disparate deployments (e.g. New York City and rural Kansas) will be just as useful in say Seoul and Jakarta as well as remote African, central Asian, and South American locales.

 

[nutzo]

I disagree. Low orbit satellite communication constellations are going to make 5G irrelevant.
They are cheaper and cost less to keep running than the equivalent terrestrial capacity.

We'll see cell towers being taken down en masse before the end of the 2020's, with only regulatory inertia (i.e., preserving the ability of old devices to make phone calls) as a barrier to it.

Even the low obit satellites have significantly worse latency than a cell tower.

The low obit satellites will be more useful in delivering decent internet coverage to more remote location. Locations that don't have the infrastructure to support 5G cell towers.

 

[pcgeekesq]

[and if you've tried to get a GPS lock on a device that lacks any 'helper' radios i.e. AGPS inside a concrete jungle or other interference-ridden environment, you understand why relying solely on low-earth-orbit infrastructure isn't a great idea]

GPS has a 31 or so satellites, and it's common enough for most of them to be obscured.
Musk is going to put up thousands of satellites. That'll will make a big difference.

 

[pcgeekesq]

Even the low obit satellites have significantly worse latency than a cell tower.

"SpaceX expects its own latencies to be between 25 and 35ms, similar to the latencies measured for wired Internet services." -- https://arstechnica.com/information...lite-internet-with-low-latency-gigabit-speed/

For most people (who are not high-speed traders or pro eSport competitors) that's low enough.

 

[IdiotInCharge]

GPS has a 31 or so satellites, and it's common enough for most of them to be obscured.
Musk is going to put up thousands of satellites. That'll will make a big difference.

I don't disagree from a field-of-view perspective, but rather just wanted to point out that regardless of the number of satellites in use, atmospheric disturbances like thunderstorms and upper-atmospheric disturbances from solar winds and solar flare activity are still going to affect reliability, and affect mobile terrestrial radios more so than fixed groundstations that can be tuned for the network and perhaps have fewer limitations in terms of power.

"SpaceX expects its own latencies to be between 25 and 35ms, similar to the latencies measured for wired Internet services." -- https://arstechnica.com/information...lite-internet-with-low-latency-gigabit-speed/

For most people (who are not high-speed traders or pro eSport competitors) that's low enough.

That's absolutely perfect so long as those latencies are kept in check when the network is stressed .

[also important to determine whether they're talking between comsat and mobile radio, or if that latency is mobile radio -> comsat network (could hit more than one sat) -> ground station -> internet service (say Google) -> ground station -> comsat network -> mobile radio; of concern is the 600ms latency quoted for geosynch comsats, which sounds like about half what they get to actual WAN services]

 

[pcgeekesq]

That's absolutely perfect so long as those latencies are kept in check when the network is stressed .

[also important to determine whether they're talking between comsat and mobile radio, or if that latency is mobile radio -> comsat network (could hit more than one sat) -> ground station -> internet service (say Google) -> ground station -> comsat network -> mobile radio; of concern is the 600ms latency quoted for geosynch comsats, which sounds like about half what they get to actual WAN services]

I believe I read SpaceX anticipates 25-35ms user-pings-server latencies, like you'd measure at speedtest.net.

Time-of-flight latency between a radio on the ground and a LEO satellite 750 miles up is 4 ms each way.
In contrast, the same metric for GEO sats (as used today) at 22,0000 miles is 118 ms each way.
Note that GPS sats orbit at 12,000 miles up -- 16 times farther away (and therefore with 256x the signal loss) than SpaceX's sats.

And SpaceX plans to route data sat-to-sat (using lasers I think) before downlinking to earth. Light is fast, so delays in the router hardware will probably dominate, but those delays should be comparable to the router delays we see on terrestrial systems.

 

[travisty]

[also important to determine whether they're talking between comsat and mobile radio, or if that latency is mobile radio -> comsat network (could hit more than one sat) -> ground station -> internet service (say Google) -> ground station -> comsat network -> mobile radio; of concern is the 600ms latency quoted for geosynch comsats, which sounds like about half what they get to actual WAN services]

Nah they're talking about predictive connections. Each node (sat/ground station) predicts what will come in and respond accordingly. It's magic!


Seriously though I am going to be amazed if the true round trip latency isn't at least 100 (still goo enough for most cases). I'd love to be surprised though!

 

[pcgeekesq]

Seriously though I am going to be amazed if the true round trip latency isn't at least 150. I'd love to be surprised though!

Why ? Do the math for LEO: 750 miles up, 4ms. 1500 miles sideways because you're communicating from (say) NYC to Austin, 8 ms. 750 miles down, 4ms. That's 16 ms of time-in-flight delays, half of it unavoidable even if you could run a laser in a vacuum tube from NYC to Austin. Add router delays, which might be fewer than in a terrestrial network, and it ought to be similar to what good broadband gets you now.

 

[Zarathustra[H]]

I can't help but wonder what the internet landscape will look like 5 years from now.

Is home land based internet going to be a thing of the past? Will 5G replace it all, with every device having its own 5G sim card, killing off the local network all together outside the business community?

I kind of like having my own LAN behind my own firewall.

 

[Nukester]

I can't help but wonder what the internet landscape will look like 5 years from now.

Is home land based internet going to be a thing of the past? Will 5G replace it all, with every device having its own 5G sim card, killing off the local network all together outside the business community?

I kind of like having my own LAN behind my own firewall.

Simple answer... NO. 5 years isn't long enough for such a drastic change. Maybe 20 years.. Things move slow with such a large country as ours. Infrastructure, infrastructure, infrastructure. I remember when I was young, I thought a lot would happen in 5 years. Boy have I learned. I'm still not in my flying car.

 

[DukenukemX]

Yet still with 3gb data caps in 2035.

 

[IdiotInCharge]

I'm still not in my flying car.

Or self-driving car...

With most massive leaps in realized technology, there has been a driver of some sort; we're not getting self-driving cars until the cars can actually do that, but once they can, they'll be everywhere; same for mass-market flying cars, which will also be self-driving.

 

[IdiotInCharge]

I kind of like having my own LAN behind my own firewall.

You can still have your own LAN behind your own firewall without wires. Perhaps less easy to control, but that's all going to come down to how secure the controllers can be made. Microsoft's push for IoT security is an indicator that I'm watching here.

 

[Zarathustra[H]]

Simple answer... NO. 5 years isn't long enough for such a drastic change. Maybe 20 years.. Things move slow with such a large country as ours. Infrastructure, infrastructure, infrastructure. I remember when I was young, I thought a lot would happen in 5 years. Boy have I learned. I'm still not in my flying car.

Fair,

But with wireless technologies it is much easier to distribute infrastructure. You are not having to run a wire to every house.

You ave existing towers, you just need to install 5G transmitters/receivers at the top of them, and - where necessary - snake new wired infrastructure between them in order to increase their bandwidth capacity.

It's not like building out completely new infrastructure.

 

[Ocellaris]

Why ? Do the math for LEO: 750 miles up, 4ms. 1500 miles sideways because you're communicating from (say) NYC to Austin, 8 ms. 750 miles down, 4ms. That's 16 ms of time-in-flight delays, half of it unavoidable even if you could run a laser in a vacuum tube from NYC to Austin. Add router delays, which might be fewer than in a terrestrial network, and it ought to be similar to what good broadband gets you now.

5G is for delivering fast connections over short distances. I think you are underestimating the amount antenna design and power needed to transmit from a mobile device to a satellite, making it impractical for handheld use on a large scale... If you think ground networks can get congested now, imagine if everyone had a cell phone which could transmit to satellites hundreds of miles away.

Antenna designs for satcom are also significantly larger when the device needs to transmit and received. This isn’t an issue that will get solved in the next 10 years to the point where it will fit inside a modern phone.

 

[oROEchimaru]

whats up with the 5g condom graphic in the background of the video?

also 5g is going to blow if it really consumes 20-50% more power (new moto phone 5g adapter has a 2000 battery pack incuded)... i dont need "Faster speeds" i need more reliable connection, more affordable and less power consumption. However if "faster = less congestion" that is great. it is going to be dumb to stream a 4k video with a 3g cap lol

 

[pcgeekesq]

I think you are underestimating the amount antenna design and power needed to transmit from a mobile device to a satellite, making it impractical for handheld use on a large scale...
Antenna designs for satcom are also significantly larger when the device needs to transmit and received. This isn’t an issue that will get solved in the next 10 years to the point where it will fit inside a modern phone.

From the 2012 FAQ: "The Thuraya XT weighs 193 grams [7 ounces] and measures 128x53x26.5mm (l x w x d)." [aprox. 5" x 2" x 1"]
-- Thuraya XT Satellite Phone, World Smallest Satellite Phone

And that's for a 2012 phone that talks to a GEO satellite stationed over Asia -- 30 times as far away as SpaceX's LEO sats, with 900x the channel loss.

Multi-thousand-satellite LEO comm systems are a whole different beast than GEO.

 

[IdiotInCharge]

whats up with the 5g condom graphic in the background of the video?

also 5g is going to blow if it really consumes 20-50% more power (new moto phone 5g adapter has a 2000 battery pack incuded)... i dont need "Faster speeds" i need more reliable connection, more affordable and less power consumption. However if "faster = less congestion" that is great. it is going to be dumb to stream a 4k video with a 3g cap lol

One trend has been increases in power usage per transmission time, but with such increases in transmission bandwidth that needed transmission time per application transaction goes down, while at the same time number of users serviceable per channel increases.

 

[MMitch]

From the 2012 FAQ: "The Thuraya XT weighs 193 grams [7 ounces] and measures 128x53x26.5mm (l x w x d)." [aprox. 5" x 2" x 1"]
-- Thuraya XT Satellite Phone, World Smallest Satellite Phone

And that's for a 2012 phone that talks to a GEO satellite stationed over Asia -- 30 times as far away as SpaceX's LEO sats, with 900x the channel loss.

Multi-thousand-satellite LEO comm systems are a whole different beast than GEO.

He's right (The person you replied to), 5G has way more use than phones. Think about self-driving car connected together, AR/VR, etc... The key advantage is shorter distance with higher bandwidths thru all devices.
From what I can recall from my short talk with our application engineer, there will be 2 frequencies; one low, typical cellphone and one high (~67GHz from memory?) which will deliver ton of data in very low latency but will be highly directional (AR/VR headset, iot devices, etc)

5G has more applications possibility than we think.

 

[Zarathustra[H]]

I've started reading about 5G technology, and my understanding is as follows:

There are two frequency ranges.

Frequency Range 1 (~600Mhz - ~6ghz):

This is where current mobile networks tend to communicate. Here your max band width is going to be 100Mhz, thus limiting it to speeds between 15%-50% faster than LTE.

The big promise of 5G seems to come in Frequency Range 2 which operates at 24Ghz-86Ghz. Here you can use 400mhz bandwidth, and get all the promises of 5G.

The problem is this. This is solidly in the millimeter wave range. The higher the frequency/shorter wavelength range drops off considerably, and you also can't penetrate walls as well. This is why low frequency (650Mhz) tend to have better signal strength indoors.

It would seem to me that to see the full benefit of high frequency 5G would require you to be within a few hundred yards of a tower, and have perfect line of sight...

Somehow 5G seems less impressive now...

 

[MMitch]

I've started reading about 5G technology, and my understanding is as follows:

There are two frequency ranges.

Frequency Range 1 (~600Mhz - ~6ghz):

This is where current mobile networks tend to communicate. Here your max band width is going to be 100Mhz, thus limiting it to speeds between 15%-50% faster than LTE.

The big promise of 5G seems to come in Frequency Range 2 which operates at 24Ghz-86Ghz. Here you can use 400mhz bandwidth, and get all th epromises of 5G.

The problem is this. This is solidly in the millimeter wave range. The higher the frequency/shorter wavelength range drops off considerably, and you also can't penetrate walls as well. This is why low frequency (650Mhz) tend to have better signal strength indoors.

It would seem to me that to see the full benefit of high frequency 5G would require you to be within a few hundred yards of a tower, and have perfect line of sight...

Somehow 5G seems less impressive now...

Well, you don't really need antenna with all the iot devices (edit: I think they want to connect devices together... I doubt it will be cellphones at those frequencies and battery would be killed)

 

[IdiotInCharge]

Somehow 5G seems less impressive now...

It's half replacing 4G and LTE and half get rid of all the wires!

Don't necessarily need USB3+, Thunderbolt, or 10Gbit ethernet or even 802.11 if your bandwidth simply scales tremendously with distance on a single radio. Get close and you're at multi-gigabit speeds, get far away and you're at 4G. For anything mobile, that'll be all you actually need, and I can see it simplifying devices where not only will it increase performance for mainstream devices but also increase portability as well as decrease unit cost for simple devices and those destined for developing markets.

 

[Deleted member 93354]

I disagree. Low orbit satellite communication constellations are going to make 5G irrelevant.
They are cheaper and cost less to keep running than the equivalent terrestrial capacity.

We'll see cell towers being taken down en masse before the end of the 2020's, with only regulatory inertia (i.e., preserving the ability of old devices to make phone calls) as a barrier to it.

Blah ha ha ha ha ha oh that's rich. There isn't near enough frequency bandwidth available for the coverage a satellite provides. Its' one of the reasons Iridium was so freaking expensive. Satellite bandwidth time is f'n expensive.

 

[pcgeekesq]

5G has more applications possibility than we think.

Perhaps. But as a cellular/broadband WAN option, it has more issues than just the wireless link to the user device.
You have to deploy an infrastructure to carry all that high-speed data from tower to tower, and terrestrial infrastructure is expense and faces lots of physicla and regulatory hurdles.

SpaceX and it's MMSL (Massively Multi-Satellite LEO) satcomm rivals don't need terrestrial infrastructure.
The satellites are the infrastructure.

 

[pcgeekesq]

Blah ha ha ha ha ha oh that's rich. There isn't near enough frequency bandwidth available for the coverage a satellite provides. Its' one of the reasons Iridium was so freaking expensive. Satellite bandwidth time is f'n expensive.

Iridium originally had 77 sats, although they've been putting up more recently for Iridium NEXT.
There are thousands of sats in a MMSL constellation, each with it's own footprint. That provides a huge amount of system bandwidth.

 

[Deleted member 93354]

I've started reading about 5G technology, and my understanding is as follows:

There are two frequency ranges.

Frequency Range 1 (~600Mhz - ~6ghz):

This is where current mobile networks tend to communicate. Here your max band width is going to be 100Mhz, thus limiting it to speeds between 15%-50% faster than LTE.

The big promise of 5G seems to come in Frequency Range 2 which operates at 24Ghz-86Ghz. Here you can use 400mhz bandwidth, and get all the promises of 5G.

The problem is this. This is solidly in the millimeter wave range. The higher the frequency/shorter wavelength range drops off considerably, and you also can't penetrate walls as well. This is why low frequency (650Mhz) tend to have better signal strength indoors.

It would seem to me that to see the full benefit of high frequency 5G would require you to be within a few hundred yards of a tower, and have perfect line of sight...

Somehow 5G seems less impressive now...

Pretty much this. This is why Sprint can't do data and voice at the same time. They use a lower frequency, so the bandwidth isn't there to support both. But it allows Sprint to build fewer towers.

 

[Deleted member 93354]

Iridium originally had 77 sats, although they've been putting up more recently for Iridium NEXT.
There are thousands of sats in a MMSL constellation, each with it's own footprint. That provides a huge amount of system bandwidth.

I would @#$@# my pants if they managed to stay out of each other's hair (Delivery cone) from LEO. But that's still considerably more area than a traditional cell tower. And no matter which way you slice it, LEO's cost more to operate because they have a short life span (relatively speaking). They burn fuel quicker and reenter orbit quicker.

 

[BoogerBomb]

American companies have ZERO reason to upgrade to 5G. They rake in tens of billions in profits off the shitty LTE we have now with stupid low data caps on efverything rendering them barely useful. China has the money and ability to outbuild the U.S. by a factor of a1000 I imagine just to make sure that they beat us to it and I have belief that they will have widespread 5G to consumers years before its even beyond testing in the U.S. There is too much lobbying against consumers in the U.S. for any new technology to realistically see the light of day any time soon.

 

[Zarathustra[H]]

It's half replacing 4G and LTE and half get rid of all the wires!

I like wires though. I can be reasonably certain they are not being intercepted by anyone other than myself, because I control both sides of them, and I never have to worry about RF interference or other connectivity problems. Plug them in, and they work.

The irrational hate of wires creates more problems, IMHO.

 

[Ocellaris]

From the 2012 FAQ: "The Thuraya XT weighs 193 grams [7 ounces] and measures 128x53x26.5mm (l x w x d)." [aprox. 5" x 2" x 1"]
-- Thuraya XT Satellite Phone, World Smallest Satellite Phone

And that's for a 2012 phone that talks to a GEO satellite stationed over Asia -- 30 times as far away as SpaceX's LEO sats, with 900x the channel loss.

Multi-thousand-satellite LEO comm systems are a whole different beast than GEO.

That antenna is fucking huge compared to what’s in a cell phone and how it needs to be packaged.

 

[BoogerBomb]

My Note 4's antenna cant even pick up a full fucking signal even when Im standing right under a god damn cell tower.

 

[pcgeekesq]

That antenna is fucking huge compared to what’s in a cell phone and how it needs to be packaged.

You kids don't remember how big cell phones used to be.
But yeah, it's big compared to a small modern phone.
But it's old tech for communicating with a much-farther-away satellite than SpaceX will be using.

 

[pcgeekesq]

I like wires though.

Me too. If it isn't mobile, it doesn't use a mobile connection.

 

[pcgeekesq]

I would @#$@# my pants if they managed to stay out of each other's hair (Delivery cone) from LEO. But that's still considerably more area than a traditional cell tower. And no matter which way you slice it, LEO's cost more to operate because they have a short life span (relatively speaking). They burn fuel quicker and reenter orbit quicker.

People are projecting that mass-produced MMSL sats may cost less than $1 million each, built and launched. That's on par with the cost of a cell tower. And you don't have to pay rent on them like you do a cell tower, weather/vandals can't damage them, and so on.

A lot of people have looked in to this, really looked at it, hired engineers, spent money, built prototype sats and launched them, all that.
They've concluded it's the future of WAN communication. I think they're right.

 

[Deepminer]

Is it just me, or does that stock photo seem like a promo for 5g condoms? Is that how they're going to get 5G to broadcast everywhere? lol

 

[cageymaru]

It's not a condom. Ha ha ha. I had to do a double take after I posted the article because I thought it was one also. I lost my voice laughing so hard today. Ha ha ha ha ha!

 

[nutzo]

It's half replacing 4G and LTE and half get rid of all the wires!

Don't necessarily need USB3+, Thunderbolt, or 10Gbit ethernet or even 802.11 if your bandwidth simply scales tremendously with distance on a single radio. Get close and you're at multi-gigabit speeds, get far away and you're at 4G. For anything mobile, that'll be all you actually need, and I can see it simplifying devices where not only will it increase performance for mainstream devices but also increase portability as well as decrease unit cost for simple devices and those destined for developing markets.

If you need to transfer large amount of data, wireless will never come close to a wire connection.
Wireless is shared, Wired is not (if using switches).

When I run updates while setting up a new laptop, I'll plug it into a 1gb wired connection. Wireless takes several times longer to download the updates from my server.

When we setup equipment for a training class, and need to copy 40GB's to each training machine, even 1gb ethernet is too slow. We use an external USB 3.0 SSD drive.