Uncapped wireless vs fibre - Bandwidth comparison

[Geoff.D]

This here. We keep hearing about how spectrum is a problem when there are already solutions. When they eventually get the spectrum they'll complain they don't have the *base stations (*Geoff doesn't seem to like towers).


And one solution to that is to ensure two neighbouring cells don't use the same frequencies. Another is to reduce the power and range and then put cells closer together. But networks try to take shortcuts by using as much spectrum and power as possible which is the wrong approach. While this doesn't change we will keep having a spectrum shortage because it's being wasted.

Also Shannon's limit assumes that spectrum is used as a shared resource. Newer technologies take advantage of beam forming or interference patterns so two users can be using the same frequency.

A classic example of a little knowledge being more dangerous than no knowledge.

 

[Swa]

A classic example of a little knowledge being more dangerous than no knowledge.

What do you mean?

 

[Johand]

This here. We keep hearing about how spectrum is a problem when there are already solutions. When they eventually get the spectrum they'll complain they don't have the *base stations (*Geoff doesn't seem to like towers).

Spectrum is absolute. You can use more base stations, with less power but that messes with your S/N ratio which means less channel capacity (S reduces due to decreased power, N increases due to neighbouring cells and devices on that cell) . Anyway you need to run fiber to every base station so at some point you are pretty much running fiber past all the houses anyway.

And one solution to that is to ensure two neighbouring cells don't use the same frequencies. Another is to reduce the power and range and then put cells closer together. But networks try to take shortcuts by using as much spectrum and power as possible which is the wrong approach. While this doesn't change we will keep having a spectrum shortage because it's being wasted.

Errr....network engineering is more complex than that. Factors like elevation, buildings, vegetation, available space for base stations makes this a complex problem. I had terrible reception at my house because conspiracy nuts kept blocking different sites (health, property values) through objections during planning process - it took networks longer than 5 years to get approval for a site. They had to jack up the power of the surrounding cells to get any reception in our area... And that obviously rippled through their entire network plan. Just erecting more base stations IS difficult, getting them to ideal locations is more difficult. It is not just about the money.

Also Shannon's limit assumes that spectrum is used as a shared resource. Newer technologies take advantage of beam forming or interference patterns so two users can be using the same frequency.

Shannon's limit is the ultimate theoretical limit on channel capacity. There is no "technology" around it. Spread spectrum use overlapping frequencies but that increases noise. Beam forming is just a way to divide a cell into sectors - helps a bit but same fundamental problems as additional cells, and can't magic more capacity (mostly used to reduce interference between neighboring cells, allowing you to jack up your power S) . Interference patterns does not do anything to Shannon's limit.

 

[Swa]

Spectrum is absolute. You can use more base stations, with less power but that messes with your S/N ratio which means less channel capacity (S reduces due to decreased power, N increases due to neighbouring cells and devices on that cell) . Anyway you need to run fiber to every base station so at some point you are pretty much running fiber past all the houses anyway.

It may be, but either way the issue in most cases is a lack of proper coverage. Networks complain about a lack of spectrum but they don't have the base stations to support it so this will simply be the new excuse and it's precisely because they've not used it efficiently and try to turn power to the max instead of focusing on the fundamental infrastructure.

Errr....network engineering is more complex than that. Factors like elevation, buildings, vegetation, available space for base stations makes this a complex problem. I had terrible reception at my house because conspiracy nuts kept blocking different sites (health, property values) through objections during planning process - it took networks longer than 5 years to get approval for a site. They had to jack up the power of the surrounding cells to get any reception in our area... And that obviously rippled through their entire network plan. Just erecting more base stations IS difficult, getting them to ideal locations is more difficult. It is not just about the money.

I never said it's not but there's no denying that base stations are too sparse and this IS because of a lack of investment. I also know a thing about bad reception. The Vodacom tower is within walking distance from me yet I have to stand in a specific spot to get coverage. If terrain is an issue there should be more in locations so this issue doesn't exist. It's also not the only Vodacom tower where I've discovered weird quirks.

Also when it comes to coverage the nearest shouldn't give you just too bars so that if that goes there's no signal at all.

Shannon's limit is the ultimate theoretical limit on channel capacity. There is no "technology" around it. Spread spectrum use overlapping frequencies but that increases noise. Beam forming is just a way to divide a cell into sectors - helps a bit but same fundamental problems as additional cells, and can't magic more capacity (mostly used to reduce interference between neighboring cells, allowing you to jack up your power S) . Interference patterns does not do anything to Shannon's limit.

That is actually not the case. Newer technology being developed isn't just dividing a cell into sectors but giving a separate sector to each device, essentially using the same frequency for more than one user. Same with interference patterns used to essentially create a cell around a person.

This wan't available then when bandwidth wasn't even a concept. One thing that's remained constant through history is that whenever someone comes up with a limit someone else comes with a way around it.

 

[cavedog]

More towers are just not financially viable when you factor in the electricity usage or lack of reliable electricity. Rent and backhaul to the tower and then get 5 x R250 uncapped users saturating the towers capacity 80% of the time. What kind of return is that on your investment.

Not to mention the ill informed community that will burn down a funny looking palm tree because it looks like 5G.

It's a simple as mobile network will just not be able to compete. A 4k movie of decent quality will eat ~50GB to 60GB easy. 10 households streaming a movie in prime time at 6pm kills the tower for everyone.

 

[microsoftword]

more base tower station things = more shit taking up physical space

why not allocate unused radio space? itll give us more physical space and money to do things

 

[microsoftword]

and if you have more tower base things station mabobs youre running fibre to more places so like less incentive for those fibre covered areas to use radio

 

[Geoff.D]

It may be, but either way the issue in most cases is a lack of proper coverage. Networks complain about a lack of spectrum but they don't have the base stations to support it so this will simply be the new excuse and it's precisely because they've not used it efficiently and try to turn power to the max instead of focusing on the fundamental infrastructure.

Another very important aspect of understanding the work done by Shannnon, Hartley and Nyquist is to truly understand the power-bandwidth tradeoffs. Only people who do not understand this make statements like you make here.

That is actually not the case. Newer technology being developed isn't just dividing a cell into sectors but giving a separate sector to each device, essentially using the same frequency for more than one user. Same with interference patterns used to essentially create a cell around a person.

This wan't available then when bandwidth wasn't even a concept. One thing that's remained constant through history is that whenever someone comes up with a limit someone else comes with a way around it.

All this "Stuff" about cell structures and patterns etc, is only about attempting to improve spectrum reuse, NOT about trying to find a way around the Shannon limits.

The very fact that you make a statement such as "when bandwidth wasn't even a concept" proves you know next to nothing about electromagnetic wave theory!

All the way back to Maxwell, that concept existed and was understood.
and BTW, no one has yet found a way to exceed the limits as defined by the Hartley-Shannon limits. Modern technologies are getting close to it and are slowly approaching that limit, but until plenty about what we know about electromagnetic wave theory collapses, there is at the moment no chance in hell that anyone will exceed or even equal those limits.
You cannot possibly confuse the basics of information theory with the techniques used to try to improve on spectrum usage with fancy antennas, beamforming and all that stuff. The Hartley-Shannon limits still end up being the limiting factor.

It is nigh on impossible ever to locate base stations at the exact optimum point in whatever cellular reuse pattern you choose to use. Even in straight forward " flat-earth simulations" its virtually impossible, let alone when you take natural topography and earth curvature into account. And then factor in natural vegetation and finally add on top of that the problems associated with man-made structures, and you very quickly get to a point where optimum coverage is a pipe dream.

 

[Swa]

More towers are just not financially viable when you factor in the electricity usage or lack of reliable electricity. Rent and backhaul to the tower and then get 5 x R250 uncapped users saturating the towers capacity 80% of the time. What kind of return is that on your investment.

Not to mention the ill informed community that will burn down a funny looking palm tree because it looks like 5G.

It's a simple as mobile network will just not be able to compete. A 4k movie of decent quality will eat ~50GB to 60GB easy. 10 households streaming a movie in prime time at 6pm kills the tower for everyone.

But adding spectrum is so that there's areas between where the spectrum isn't used? This reminds of how they kept handing out numbers till Icasa eventually said sorry no more coming and you have to use what you have efficiently. That's partly what caused the problem where MTN will cancel a sim while you are literally using it.

Anyway I do not believe that Rain has all of that spectrum saturated.

 

[microsoftword]

@Geoff.D you sound like you know your stuff i have a genuine question: is the Gov not allocating spectrum because of incompetence or is there a technical reason you can share so I can improve my understanding

 

[Geoff.D]

More towers are just not financially viable when you factor in the electricity usage or lack of reliable electricity. Rent and backhaul to the tower and then get 5 x R250 uncapped users saturating the towers capacity 80% of the time. What kind of return is that on your investment.

Not to mention the ill informed community that will burn down a funny looking palm tree because it looks like 5G.

It's a simple as mobile network will just not be able to compete. A 4k movie of decent quality will eat ~50GB to 60GB easy. 10 households streaming a movie in prime time at 6pm kills the tower for everyone.

Kills the base station or network node, the tower is just the structure on which all the "stuff" is mounted.

Your point is correct about the financial side -- it is extremely difficult to make these pipe dream radio networks financially viable and keep the service offered affordable.

Without sufficient backhaul capacity to cater for peak demand, users on a mobile network will always suffer. And then you have to ask yourself that IF you are going to deploy a fibre network to feed those base stations then why not just give all fixed location users direct access to that capacity and let them deal with the last inch problem to their devices if they insist on being tether-free? That will leave the wireless technologies free to use the available capacity for truly mobile applications instead?

 

[Geoff.D]

@Geoff.D you sound like you know your stuff i have a genuine question: is the Gov not allocating spectrum because of incompetence or is there a technical reason you can share so I can improve my understanding

Unfortunately, the situation is SA is all about corruption and political interference in a highly complex technical environment by people who have no understanding about the damage they are doing.

Then add to that gross incompetence within the DoC, plus a paucity of technical expertise within ICASA.

But at the same time, as SWA points out, most of the current operators are also extremely lax about exploiting the capacity they already have at their disposal.
(Failure to migrate to newer technologies and re-farm their existing capacity, undertake much better coverage planning, etc).

The ITU lays out the spectrum usage plans up to 30 years ahead, develops the band plans, lays down the broad interference mitigation strategies and sharing strategies and then each country has to implement the same within fixed time frames within their territories.
The current "5G" band plans were finalised within the ITU as far back as WRC 2000 ( 20 years ago!). And SA was directly involved in that process, so we cannot claim we did not know what was coming or what was being prepared for.

And that is where our lot have completely lost the plot. The delays in sorting out even a simple thing such as the DTT migration causes a ripple effect throughout the entire region. The precise reason for the freeing up of low-frequency bands is to allow these to be primarily used in sparsely populated areas for delivering the sort of services us urban dwellers increasingly take for granted, where pervasive fibre networks are still very difficult to implement and afford.

Spectrum not utilised is spectrum wasted forever. The negative impact on our economic growth is there for all to see.

 

[Swa]

Another very important aspect of understanding the work done by Shannnon, Hartley and Nyquist is to truly understand the power-bandwidth tradeoffs. Only people who do not understand this make statements like you make here.

No that is not the issue. The problem is that the operators have not tried to use spectrum efficiently. Instead of trying to provide as much coverage as possible using a specific band they've instead tried to solve the problem using more spectrum. So essentially erecting another adjacent tower on different bands and then cranking up the power to extend the edge to cover more ground.

Increasing the power to improve signal to noise isn't the answer as it simply increases the distance it causes interference as well. Instead what you should do is decrease it as much as possible and then space cells closer together to decrease the coverage gap. Then essentially build another network on different spectrum between these cells to cover any coverage gaps.

My issue isn't a lack of understanding these concepts but a problem with how it's implemented.

All this "Stuff" about cell structures and patterns etc, is only about attempting to improve spectrum reuse, NOT about trying to find a way around the Shannon limits.

The very fact that you make a statement such as "when bandwidth wasn't even a concept" proves you know next to nothing about electromagnetic wave theory!

All the way back to Maxwell, that concept existed and was understood.
and BTW, no one has yet found a way to exceed the limits as defined by the Hartley-Shannon limits. Modern technologies are getting close to it and are slowly approaching that limit, but until plenty about what we know about electromagnetic wave theory collapses, there is at the moment no chance in hell that anyone will exceed or even equal those limits.
You cannot possibly confuse the basics of information theory with the techniques used to try to improve on spectrum usage with fancy antennas, beamforming and all that stuff. The Hartley-Shannon limits still end up being the limiting factor.

It is nigh on impossible ever to locate base stations at the exact optimum point in whatever cellular reuse pattern you choose to use. Even in straight forward " flat-earth simulations its virtually impossible, let alone when you take natural topography and earth curvature into account. And the factor in naturl vegetation and finally add on top of that the problems associated of man-made structures, and you very quickly get to a point where optimum coverage is a pipe dream.

As I said the modern concept of bandwidth was just a dream. The closest thing at the time was radio broadcasts and the telegraph. So I still have my doubt that the limit will prove to be a real limit as all limits people have come up with so far have been overcome.

But that is not what modern technologies focus on. Instead of overcoming it they circumvent it so it doesn't apply. Actually things like beam forming and interference patterns don't do it justice as this is so far beyond that it pales in comparison. Essentially what it does is to reuse the same bandwidth or form cells where users are. It's not improving spectrum usage but more like adding another line to a user when a single line is shared by two users.

 

[Geoff.D]

So essentially erecting another adjacent tower on different bands and then cranking up the power to extend the edge to cover more ground.

In other words the power - bandwidth tradeoffs!

As I said the modern concept of bandwidth was just a dream.

No, it was there for all to see and use. BUT, the technologies were not available to exploit it! They still had to be developed. The basics of how it all would work were already there! This is precisely where the power of Moore's Law comes from! Where one assumes that whatever technical constraint currently exists will be solved by the time you need it to be solved. It frees one up from being paralysed by "current technological constraints". That is the only way the ITU is able to plan 30 to 50 years ahead in this game.

Do some real reading into the subject.

 

[Swa]

In other words the power - bandwidth tradeoffs!

The wrong tradeoff, from a time when quickly covering the population was more important than using spectrum efficiently for maximum benefit. You know a large part of the 3G/LTE network was never designed? It was simply tacked onto existing 2G towers.

No, it was there for all to see and use. BUT, the technologies were not available to exploit it! They still had to be developed. The basics of how it all would work were already there! This is precisely where the power of Moore's Law comes from! Where one assumes that whatever technical constraint currently exists will be solved by the time you need it to be solved. It frees one up from being paralysed by "current technological constraints". That is the only way the ITU is able to plan 30 to 50 years ahead in this game.

Do some real reading into the subject.

Not available means it couldn't really have been understood. There has been literally a mountain of development since then such as modulation schemes and error correction to name just a bit.

 

[Geoff.D]

The wrong tradeoff, from a time when quickly covering the population was more important than using spectrum efficiently for maximum benefit. You know a large part of the 3G/LTE network was never designed? It was simply tacked onto existing 2G towers.

I am more aware of that than most because I and many others of my generation were involved in the planning of the original networks! It is appalling that the operators all deliberately chose to simply dump 3G etc onto the exiting network structures. Hence my comment supporting your view about how lax the new generation of mobile network planners are.

Not available means it couldn't really have been understood. There has been literally a mountain of development since then such as modulation schemes and error correction to name just a bit.

Now you are being willfully ignorant! The driving force that led to the development of all those modulation schemes and error corrections and all of that is precisely because the developers wanted to reach the Shannon limit. Hence they knew exactly where they wanted to be and why, and understood what needed to be done and worked out how to do it.

 

[microsoftword]

Unfortunately, the situation is SA is all about corruption and political interference in a highly complex technical environment by people who have no understanding about the damage they are doing.

Then add to that gross incompetence within the DoC, plus a paucity of technical expertise within ICASA.

But at the same time, as SWA points out, most of the current operators are also extremely lax about exploiting the capacity they already have at their disposal.
(Failure to migrate to newer technologies and re-farm their existing capacity, undertake much better coverage planning, etc).

The ITU lays out the spectrum usage plans up to 30 years ahead, develops the band plans, lays down the broad interference mitigation strategies and sharing strategies and then each country has to implement the same within fixed time frames within their territories.
The current "5G" band plans were finalised within the ITU as far back as WRC 2000 ( 20 years ago!). And SA was directly involved in that process, so we cannot claim we did not know what was coming or what was being prepared for.

And that is where our lot have completely lost the plot. The delays in sorting out even a simple thing such as the DTT migration causes a ripple effect throughout the entire region. The precise reason for the freeing up of low-frequency bands is to allow these to be primarily used in sparsely populated areas for delivering the sort of services us urban dwellers increasingly take for granted, where pervasive fibre networks are still very difficult to implement and afford.

Spectrum not utilised is spectrum wasted forever. The negative impact on our economic growth is there for all to see.

thank you for the very detailed explanation, amazing contribution to the forum

 

[Johand]

This wan't available then when bandwidth wasn't even a concept. One thing that's remained constant through history is that whenever someone comes up with a limit someone else comes with a way around it.

Shannon's limit is an absolute limit and always apply. "interference patterns" aka Beam Steering is just more spatial multiplexing and reducing multipath interference. These techniques allows us to use spectrum more efficiently but doesn't *give* us more. We can maybe get several orders of magnitude better use of spectrum, but it can't bring us close to the raw bandwidth of fiber. Even with that being said that most of network design comes down to cost efficiency - both core network and client devices. If we could switch off 2G and 3G we would have a bunch more spectrum. If we could build network towers on every street corner we would have amazing spectrum use. It would just not be cost effective - mobile networks are capital expensive. Big companies like network operators are actually in the business of capital allocation to get good returns. If they can't allocate capital effectively compared to other industries then investors don't invest.

BTW The Shannon limit isnt a physics limit - it is a mathematical limit with mathematical proof. It does not depend on the medium - it gives valid results regardless of whether you use radio or waterwaves or a string between two cans.

 

[Swa]

Shannon's limit is an absolute limit and always apply. "interference patterns" aka Beam Steering is just more spatial multiplexing and reducing multipath interference. These techniques allows us to use spectrum more efficiently but doesn't *give* us more. We can maybe get several orders of magnitude better use of spectrum, but it can't bring us close to the raw bandwidth of fiber. Even with that being said that most of network design comes down to cost efficiency - both core network and client devices. If we could switch off 2G and 3G we would have a bunch more spectrum. If we could build network towers on every street corner we would have amazing spectrum use. It would just not be cost effective - mobile networks are capital expensive. Big companies like network operators are actually in the business of capital allocation to get good returns. If they can't allocate capital effectively compared to other industries then investors don't invest.

BTW The Shannon limit isnt a physics limit - it is a mathematical limit with mathematical proof. It does not depend on the medium - it gives valid results regardless of whether you use radio or waterwaves or a string between two cans.

We can't know that. There's a saying never say never. What amazes me is how ignorant humanity has always been about its own knowledge. Even the speed of light has been broken. You can look that up if you don't believe it, it's just not been put to practical use.

What you say here is exactly what the techniques I mentioned do. Again it does not create extra spectrum but uses the spectrum multiple times. So like putting in a second line for two people that shared one. It's no different than one person using it in Cape Town and another in Johannesburg except it's on a single tower.

 

[Ockie]

I donk think the term "Wireless" should be used for this article. Rather use LTE as Wireless often refers to fixed services where they put the little dish on your roof. Two very different things.