Fibre versus Starlink Discussion

[bkp_duke]

Sure, but Starlink also isn't the shorted path as the satellites are in orbit, meaning the signal has to go up 340km and then down 340km, or further if the satellite isn't directly over both the user and the base station. 680km combined before you add any lateral distance.

Light travels 2X as fast in a vacuum as it does through glass (fiber).

Also, you apparently are making the erroneous assumption that fiber is akin to a point to point protocol. As others here have pointed out, far from it. Perform a tracert (traceroute in Linux) to 8.8.8.8 (Google DNS). You will be surprised how many hops you make to a close DNS server.

 

[banned-66611]

Light travels 2X as fast in a vacuum as it does through glass (fiber).

No it doesn't.

 

[banned-66611]

I’m just a layman. But isn’t laser beam already itself confined?

No, it's in free space and diverges from the point of origin. With optics the convergence point can be controlled somewhat but not over the very large distances between satellites.

The divergence means that not all of the light reaches the receiver, so more has to be sent, i.e. the emitter has to be on for longer to emit more photons or be higher power, both of which limit the data rate. In space it also has to deal with interference from the sun, which is not shielded by the atmosphere and at times will align with the receiver.

An alternative is to use an optical modulator but then your laser has to be powered constantly, and cooled constantly.

Fibre optics use wavelength division multiplexing, essentially sending lots of different colour light down the same cable. This is well established technology now but only just starting to see major developments with free-space laser communications.

 

[ItsNotAboutTheMoney]

Amazon is testing their rival system which claims 400Mbps and "low latency". As ever the laws of physics make both claims dubious.

Amazon’s answer to SpaceX Starlink delivers 400Mbps in prototype phase

As ever, you need to understand that low latency != fast fiber latency.

It's in comparison to typical broadband services used by rural users, which is high latency over copper or very high latency geostationary satellite.

 

[banned-66611]

As ever, you need to understand that low latency != fast fiber latency.

As ever, you need to understand the digital divide.

If Starlink is mediocre by today's standards then is it really worth investing large sums of money in when it will just put people back in the same position of having extremely slow (relative to the average) broadband in a few years?

"It's better than nothing" isn't exactly aiming high.

 

[bxr140]

If Starlink is mediocre by today's standards then is it really worth investing large sums of money in when it will just put people back in the same position of having extremely slow (relative to the average) broadband in a few years?

At the risk of speaking for everyone here, it would be useful for you to provide actual supporting logic when making claims. Near as I can figure:

1. You have not established that Starlink is mediocre by today's standards. Again, it is not; the latency is more than adequate for low latency applications, and there is no evidence that latency demands in the near future will become materially more stringent. Again, the notional speed of 100mbps, even in an "up to 100mbps" scenario, is more than adequate for today's standards and there's no evidence to suggest that average demand in the near-mid future will exceed that level, and certainly not in the next 5 years, which is +/- when this current generation of Starlinks will go EOL.
2. You have avoided quantifying load/speed/demand in the near future. Again, please explain what you think average load will be in the future. It seems as though you think you <ahem> understand this subject matter thoroughly, so a parametric projection should be no problem, yeah?
3. You have avoided addressing the cost of fiber. Again, how much will it cost to run fiber to everyone in the US, or a region...or honestly, any relevant geographical metric anywhere in the world.

You seem to fully dismiss (or not understand) the intended user base for Starlink; but instead of me making further assumptions on your understanding how about you explain who's actually going to be using Starlink? Then, iffn you wouldn't mind, please re-address the above in the context of the actual Starlink user base.

(Spoiler alert, its for people with shitty or no internet)

 

[hmcgregoraz]

As ever, you need to understand the digital divide.

If Starlink is mediocre by today's standards then is it really worth investing large sums of money in when it will just put people back in the same position of having extremely slow (relative to the average) broadband in a few years?

"It's better than nothing" isn't exactly aiming high.

The beta is "better than nothing", and the reason for that is until the more recent launches are at their proper orbit, there will be times of ZERO service. The commercial service will not be "better than nothing" class, and will be quite usable.

StarLink is by no means "large sums of money" when compared to Fiber for 80% of the US, let alone 90% of the US.

StarLink is NOT meant to replace fiber where fiber is practical.

I have said many times that if you can get Fiber, or HFC (Hybrid Fiber Coax), StarLink is NOT meant for you, which is GOOD, as it will keep the amount of subscriber density DOWN, which will help StarLink perform well for those that need it.

Please DO NOT think that SpaceX and StarLink are trying to prevent Fiber deployments, with the possible exception of where VERY HEAVY government subsidies are needed. If you look at the FCC awards, Fiber, where it made even the most remote sense, Fiber got the award.

The costs and methods Telcos and Cablecos are using to deploy fiber are NOT economically viable in many many areas.

I would love to see municipal water companies deploying fiber networks when ever they are doing pipe installation and replacement, and then leasing fiber access to vendors that want to sell service, but that still only really works well for 80% of the US population.

-Harry

 

[banned-66611]

1. You have not established that Starlink is mediocre by today's standards. Again, it is not; the latency is more than adequate for low latency applications, and there is no evidence that latency demands in the near future will become materially more stringent. Again, the notional speed of 100mbps, even in an "up to 100mbps" scenario, is more than adequate for today's standards and there's no evidence to suggest that average demand in the near-mid future will exceed that level, and certainly not in the next 5 years, which is +/- when this current generation of Starlinks will go EOL.
2. You have avoided quantifying load/speed/demand in the near future. Again, please explain what you think average load will be in the future. It seems as though you think you <ahem> understand this subject matter thoroughly, so a parametric projection should be no problem, yeah?
3. You have avoided addressing the cost of fiber. Again, how much will it cost to run fiber to everyone in the US, or a region...or honestly, any relevant geographical metric anywhere in the world.

Did you read the previous posts? For your benefit, one more time:

1. Fibre is the current generation technology and will be what we use for the next century at least, with speeds continuing to increase over the same lines. The goal here is not to bring people up to the last generation, it's to bring them up to the current one AND future proof them so that in 10 years time they aren't back to square one with a slow connection again.

2. Historically bandwidth requirements have continued to rise steadily over time and it seems reasonable to think that they will continue to do so. 5 years ago we were not seeing people regularly uploading 4k video to YouTube, for example. We were not seeing game downloads nearing 100GB. 4k RDP was largely unknown, now I do it every day, on two screens.

We have seen low latency demands increase too. For interactive services like RDP and video conferencing, but also for things like games. Many games already have a "wifi filter" that blocks players with low or variable latency. Streaming game services (e.g. Sony's PS Now) are rapidly gaining popularity too.

It seems extremely unlikely that demands will not continue to increase, while at the same time Starlink's numbers continue to get worse as more subscribers are added to the service.

3. The cost of fibre is considerable up front, but so was copper telephone lines and electricity distribution. It's worth investing in though because that infrastructure will be good for at least a century, probably longer. Amortized over any reasonable estimated lifespan it's quite affordable. Much of the infrastructure, like poles, is already there.

And let's compare to the cost of Starlink. Satellites need replacing every few years.

Allow me to add a 4th point to the mix: The environmental impact. Starlink is not environmentally friendly. Disposable satellites, launched from the ground with huge energy requirements and massive emissions every time. Even the transceivers are pulling 150W. We should be building a sustainable future.

 

[bkp_duke]

(BS snipped out)

Allow me to add a 4th point to the mix: The environmental impact. Starlink is not environmentally friendly. Disposable satellites, launched from the ground with huge energy requirements and massive emissions every time. Even the transceivers are pulling 150W. We should be building a sustainable future.

Yet you completely discount that trenching fiber doesn't have a HUGE environmental impact.

Whatever chump. All your arguments are 1/2 baked, and you keep moving the goalposts when those here in the know that WORK in the industry shoot down your theories.

 

[bkp_duke]

Agreed.

Although there will be latency incurred in the sat-to-sat hops too... but I expect there may be less of those...

But there are fewer hops.

I 100% guarantee you that "crossing the pond" with Starlink will be faster than with undersea cable. The speed advantage of light in a vacuum over long distances with rapidly overcome fiber.

 

[scaesare]

But there are fewer hops.

I 100% guarantee you that "crossing the pond" with Starlink will be faster than with undersea cable. The speed advantage of light in a vacuum over long distances with rapidly overcome fiber.

Yup, that's what I was referring to when I said "...I expect there may be less of those..."

 

[hmcgregoraz]

Did you read the previous posts? For your benefit, one more time:

1. Fibre is the current generation technology and will be what we use for the next century at least, with speeds continuing to increase over the same lines. The goal here is not to bring people up to the last generation, it's to bring them up to the current one AND future proof them so that in 10 years time they aren't back to square one with a slow connection again.

2. Historically bandwidth requirements have continued to rise steadily over time and it seems reasonable to think that they will continue to do so. 5 years ago we were not seeing people regularly uploading 4k video to YouTube, for example. We were not seeing game downloads nearing 100GB. 4k RDP was largely unknown, now I do it every day, on two screens.

We have seen low latency demands increase too. For interactive services like RDP and video conferencing, but also for things like games. Many games already have a "wifi filter" that blocks players with low or variable latency. Streaming game services (e.g. Sony's PS Now) are rapidly gaining popularity too.

It seems extremely unlikely that demands will not continue to increase, while at the same time Starlink's numbers continue to get worse as more subscribers are added to the service.

3. The cost of fibre is considerable up front, but so was copper telephone lines and electricity distribution. It's worth investing in though because that infrastructure will be good for at least a century, probably longer. Amortized over any reasonable estimated lifespan it's quite affordable. Much of the infrastructure, like poles, is already there.

And let's compare to the cost of Starlink. Satellites need replacing every few years.

Allow me to add a 4th point to the mix: The environmental impact. Starlink is not environmentally friendly. Disposable satellites, launched from the ground with huge energy requirements and massive emissions every time. Even the transceivers are pulling 150W. We should be building a sustainable future.

1) No one here ever said StarLink will "replace" fiber, nor does it need to. Yes it will be "better than fiber" for some specific long distance connections in the future with ISLs, but that is a more limited application, though a very lucrative one.

2) Low latency is not what you think it is. Here are some pings to 8.8.8.8 which uses BGPAnycast to be as close as possible network wise to as many end users as possible:

CenturyLink Bonded VDSL2 (Vail AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=117 time=34.9 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=117 time=34.9 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=117 time=37.7 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=117 time=35.0 ms

CenturyLink 100/100 Small Biz GPON Fiber (Tucson AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=117 time=15.4 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=117 time=15.2 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=117 time=15.3 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=117 time=15.1 ms

Cox 100/20 Small Biz Cable (Tucson AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=117 time=26.6 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=117 time=28.4 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=117 time=24.9 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=117 time=28.0 ms

Level 3 dual bonded T1 (yes, 3/3 service, no users on at the time of the ping) (Tucson AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=115 time=17.9 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=115 time=16.4 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=115 time=16.5 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=115 time=16.3 ms

FirstDigital Business Fiber (Tucson AZ)
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=113 time=11.9 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=113 time=13.6 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=113 time=13.1 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=113 time=14.6 ms

Linode (virtual server) instance in Freemont CA
ping -c 4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=121 time=1.33 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=121 time=1.42 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=121 time=1.43 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=121 time=1.40 ms

I just downloaded 377MB on that bonded T1: Fetched 377 MB in 39min 24s (160 kB/s)

Cox has far higher latency vs the bonded T1, but I would NEVER want to be stuck on that bonded T1, ever.

3) For many many places in the US the cost of Fiber will NEVER be able to be recouped, let alone a worth while investment. I deal with some locations that do not have electrical utilities, and they never will. Currently most are Diesel based, but most will move to battery and solar.

Many places around the world are skipping wired/fiber infrastructure almost entirely, wifi, LTE, and 5G are the methods of choice for last mile due to the huge investment needed for copper or fiber, and StarLink and other LEO constellations will become the backhaul of choice.

When doing other infrastructure projects that need to be done (like water or sewer pipes), putting Fiber in makes sense. The LA department of water and power decided to do this well over a decade ago, and it has helped. https://www.ladwp.com/ladwp/faces/ladwp/partners/p-fiberoptics

4) This is BETA hardware with a BETA service, with BETA firmware. The power consumption will go down, they will start using sensors to determine when heating is needed, they will have more satellites in view, which will also reduce power consumption, etc.

The mass being launched (60 at a time on F9, far more on Starship), and the energy cost of those launches (combined) will eventually be less than the energy used to deliver the user terminals (combined) via FedEX/UPS/etc. Now think of all of the energy used for a Fiber deployment, underground or above ground, the amount of truck rolls, the amount of energy in shipping the fiber, etc.

StarLink averages this energy across a vast amount of land mass, and thus a very low energy cost per user.

-Harry

 

[ItsNotAboutTheMoney]

As ever, you need to understand the digital divide.

If Starlink is mediocre by today's standards then is it really worth investing large sums of money in when it will just put people back in the same position of having extremely slow (relative to the average) broadband in a few years?

"It's better than nothing" isn't exactly aiming high.

The digital divide is about people being able to function in a society where access to education, jobs, government services and business services requires broadband Internet good enough to do video streaming and videoconferencing.

We are doing that _now_ with ADSL, not fiber. Our pings to public DNS are at best 25ms, and regularly spike to 70ms.

The digital divide is not about the ability to work remotely in 2x 4k.
The digital divide is not about being about being able to download and play large games in minutes.

The "better than nothing beta" name reflect the new and incomplete constellation, that means that they aren't yet guaranteeing stable service, and that customers outside of a narrow latitude range have holes in coverage. It also reflects the sense of humor of a group of engineers who refer to optical inter-satellite communication as space lasers.

 

[banned-66611]

Actually more like 2/3rds.... but are you ignoring the larger point that the 150% speed advantage in a vacuum helps offset the additional distance incurred by the ground/sat round-trip?

No, scroll back and you will see I factored it in to my calculations.

For the scrolling impaired consider that the satellites are at an altitude of 320km, so that's the minimum distance that the radio signal has to travel. Of course it's usually further because they are not directly overhead. And then it has to come back down too. So immediately you have at least an extra 640km to travel.

2/3rds speed of light is a bit low for modern fibre but okay, let's go with that. Fibre can send a packet 422km in the time it takes Starlink just to go up and down, let alone sideways.

Halve it for the round trip. Most people live much closer than 210km to a CDN, so your response on fibre comes in before Starlink has even covered the distance to space and back.

 

[banned-66611]

Yet you completely discount that trenching fiber doesn't have a HUGE environmental impact.

Why would you trench it? There are already utility poles with copper telephone lines going to those locations. Or at the worst buried trunks carrying them that can be re-used.

Come on, engage your brain. Look at how other countries have done this. Japan is a great example, really difficult terrain but they were able to use what was already there.

 

[banned-66611]

2) Low latency is not what you think it is.

...

I just downloaded 377MB on that bonded T1: Fetched 377 MB in 39min 24s (160 kB/s)

Low latency is not what you think it is.

Latency is the time it takes to get a response. Say you are playing a streaming game, latency affects how long it takes the game to register your button presses and for the action you commanded to appear on your screen. Similarly if you are using remote desktop and you type sometimes latency affects how long it takes for the result of your typing to appear on screen.

By the way, fibre is much better than an ancient T1 line in this respect. As well as offer 20Gbps of bandwidth on consumer lines, ping times are extremely low as well. If you are using high bandwidth interactive services that's what you need, both of those things.

3) For many many places in the US the cost of Fiber will NEVER be able to be recouped, let alone a worth while investment.

Was the cost of installing phone lines or electricity lines recouped? Or the cost of building roads all the way out there that hardly anyone uses? What about water supply?

Maybe not, but it was worth doing anyway.

Many places around the world are skipping wired/fiber infrastructure almost entirely, wifi, LTE, and 5G are the methods of choice for last mile due to the huge investment needed for copper or fiber, and StarLink and other LEO constellations will become the backhaul of choice.

Yes, many places are short sighted and dumb.

 

[banned-66611]

The digital divide is about people being able to function in a society where access to education, jobs, government services and business services requires broadband Internet good enough to do video streaming and videoconferencing.

That's the absolute bare bones minimum today. But even now people stuck with that are being held back, cut off from some services.

For example they could apply for that remote job, but their productivity will be affected by slow broadband making their RDP session unresponsive or forcing them to wait to open up large files.

 

[bkp_duke]

Why would you trench it? There are already utility poles with copper telephone lines going to those locations. Or at the worst buried trunks carrying them that can be re-used.

Come on, engage your brain. Look at how other countries have done this. Japan is a great example, really difficult terrain but they were able to use what was already there.

Engage my brain? Says the guy that cannot defend his arguments.

MOST fiber is trenched, not up on poles. Why? Because of tornadoes, hurricanes, morons driving cars into poles, etc. you don't want the risk of a fiber break of crap strung on poles. MOST fiber in Japan is actually underground, NOT on poles, you nit wit.

I run a business that GUARANTEES 5-9s (99.999%) uptime to different datacenters on 3 continents. It takes HOURS to repair fiber (it's a bitch to splice), so no one in their right mind runs fiber on poles except to residences, where you don't have uptime guarantees.

Why don't YOU engage YOUR brain and do some research before you spout out your personal, uneducated opinions as FACT.

 

[DMC-Orangeville]

Why would you trench it? There are already utility poles with copper telephone lines going to those locations. Or at the worst buried trunks carrying them that can be re-used.

Come on, engage your brain. Look at how other countries have done this. Japan is a great example, really difficult terrain but they were able to use what was already there.

Canadian standards for this rests with the Provinces. All Provinces but Nova Scotia insist that fibre cables must be buried. We have been lobbying our Provincial Government to use overhead cabling, as most of the rural areas (like me) have their hydro delivered by utility poles. Their reply, so far: Get lost.

 

[hmcgregoraz]

Low latency is not what you think it is.

Latency is the time it takes to get a response. Say you are playing a streaming game, latency affects how long it takes the game to register your button presses and for the action you commanded to appear on your screen. Similarly if you are using remote desktop and you type sometimes latency affects how long it takes for the result of your typing to appear on screen.

By the way, fibre is much better than an ancient T1 line in this respect. As well as offer 20Gbps of bandwidth on consumer lines, ping times are extremely low as well. If you are using high bandwidth interactive services that's what you need, both of those things.



Was the cost of installing phone lines or electricity lines recouped? Or the cost of building roads all the way out there that hardly anyone uses? What about water supply?

Maybe not, but it was worth doing anyway.



Yes, many places are short sighted and dumb.

You completely ignored my statement on latency. Yes, lower latency is "better", but I would not trade the 26ms latency of the cable HFC for the 16ms latency of the dual T1 line, as the bandwidth available results in major congestion and very high latency when in use during the day.

Cox 100/20 Small Biz Cable (Tucson AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=117 time=26.6 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=117 time=28.4 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=117 time=24.9 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=117 time=28.0 ms

Level 3 dual bonded T1 (yes, 3/3 service, no users on at the time of the ping) (Tucson AZ):
ping -c4 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=115 time=17.9 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=115 time=16.4 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=115 time=16.5 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=115 time=16.3 ms

I have been involved in internet engineering for the last ~25 years, built proxy server environments, made Internet Access usable in K12 environments where it previously was not usable due to bandwidth constraints. I have deployed plenty of Point to Point wireless links due locations being unable to get effective bandwidth.

I have also been involved in enterprise networking for a considerable amount of time.

You keep missing this is NOT a one size fits all, and that there is plenty of room for StarLink.

I can even see eventually small towns will put in passive optical networks (GPON, etc), but use commercial StarLink for the uplink.

Just because we had no other choice but to dump huge amounts of $ into rural electrification and telephone (copper) networks, does not mean we should do that again with Fiber.

Plus many of those rural telephone deployments used microwave relay towers to get outside connectivity, not copper (coax or pairs) or Fiber.

Many small towns are still serviced with Geo Satellite or Microwave Relay towers for external connectivity to the PSTN, not Fiber.

We have tons of above ground pole mounted Fiber in Tucson AZ, yet the residential availability of Fiber is TINY. Still, anyone in Tucson AZ is not a primary StarLink customer, maybe out side of town in areas like Vail, but not in the city.

-Harry