4 ms·

Are you able to expand? I tend to agree but dont really know why or what / where the physical limits are. Is it the 4g etc tech that can't perform with real loa

by lugg 12y ago

Are you able to expand? I tend to agree but dont really know why or what / where the physical limits are. Is it the 4g etc tech that can't perform with real load or WiFi that will fall over? Is a mesh network an answer?

AnthonyMouse 12y ago
The problem is that wireless is a shared medium. If you have a certain amount of spectrum, you can transfer a finite number of bits per second using it, and that capacity is shared between all devices in range. The bandwidth available in a certain chunk of spectrum is a physical limit. There is no way to expand capacity other than by allocating more spectrum, and radio spectrum is extremely scarce. So the more users you have using cellular for internet, the less bandwidth they each get. The solution when you want to have "wireless" devices but there isn't enough wireless spectrum is to use wireless for the "last 50 meters" rather than the last mile. In other words, you have fiber to the premises and then WiFi within the building. By keeping the wireless power levels as low as possible you reduce the number of other users you're sharing the spectrum with and allow each user to have a satisfactory amount of bandwidth. Then reserve cellular wireless for users not in a building, e.g. accessing a map from your car, and have all manner of stupid bandwidth caps to discourage usage of cellular wireless whenever it isn't strictly necessary. > Is a mesh network an answer? No. Mesh networks would allow you to have more bandwidth per user by reducing wireless power levels and shrinking the geographic size of the collision domain, but that would require peers to act as repeaters between other peers and the tower. That's an epic fail. It adds latency, reduces battery life, impacts reliability if any peer's device is flaky, has security implications, etc. Wireless mesh networks ironically work much better for fixed-position devices than mobile devices, because they can use directional antennas, don't have the extreme power constraints mobile devices do, etc.
- lugg 12y ago
  Thanks mostly confirming what I suspected. I did think we would be able to use the spectrum more efficiently over time with better, more precise measurements but my guess is that will scale to its own limits a lot faster than a hard line will.
  - martinald 12y ago
    Much like procesing power and moore's law, we're getting very close to shannon's law (http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem. Wireless isn't this huge gamechanger that people may think. To beat shannon's law you would need so many base stations in the long term it looks awfully like what we have now with DSL/Fiber/DOCSIS and wifi APs.
- exelius 12y ago
  It all depends how much bandwidth you need for a given use case. For most use cases other than video, existing wireless technologies are sufficient and will only continue to improve as things like beam shaping and micro cells become more prevalent. Wireless is easier to deploy and as its use grows, more spectrum will likely open up. Free markets have the benefit of deploying scarce resources where they are most in demand. Implementing that spectrum is often as easy as a configuration change that can be deployed remotely. The biggest advantage of a wired home broadband connection over a wireless one that I see is the ability to stream video. A wired connection has a lot of disadvantages: it's useless outside the home and the infrastructure is significantly more expensive. Wireless has signal and coverage issues for sure, but wireless is a much more competitive industry than wired access in no small part because it's cheaper to deploy. I don't think it would be going out on a limb to say the HN audience is much wealthier and more technology-literate than the average citizen, and thus would be more likely to want multiple Internet connections and have more use cases such as running file servers, etc. I think the HN crowd overestimates what people use their Internet connections for; the average person uses their home broadband connection primarily to watch video. It's not a bad use case by any means; but I'm not sure the government needs to be building infrastructure that will be primarily used by video streaming companies. Traditional cellular wireless (which can use any mix of wireless, wired or satellite backhaul depending on the application) and Wi-Fi are flexible enough to cover the "access" part of Internet access; especially in an increasingly post-PC world.
  - AnthonyMouse 12y ago
    > For most use cases other than video, existing wireless technologies are sufficient and will only continue to improve as things like beam shaping and micro cells become more prevalent. Excluding today's most demanding use case doesn't get you anywhere. If you need a wired connection for that then you need a wired connection. The installed base of smartphones and tablets is still growing significantly and that's going to eat any technological improvements you can make to cellular and then some. Meanwhile people still want to be able to watch Netflix on a tablet from the back of a minivan, and to be able to make that happen at scale you're going to have to remove as much other traffic from the cellular network as possible. > Wireless is easier to deploy and as its use grows, more spectrum will likely open up. Free markets have the benefit of deploying scarce resources where they are most in demand. Free markets can't change the laws of physics. Allocating more spectrum is a linear increase. 20% more spectrum costs billions of dollars and only makes it 20% faster. The market solution when there is a high demand scarce resource is to shift usage to a resource that isn't as scarce for any use case where the alternative resource is viable. There is a reason one of the first things you already do when you get to a friend's house for the first time is ask for the WiFi password, and it's not because there is no cellular coverage. > The biggest advantage of a wired home broadband connection over a wireless one that I see is the ability to stream video. The advantage is that it allows you to transfer more bits cheaply. Today's most prevalent use case for that is streaming video, sure. It's hardly the only one. Anything that requires uploading or downloading large amounts of data is a bad fit for cellular. And we can't predict what interesting applications will become commonplace once the typical connection is 1Gbps. > A wired connection has a lot of disadvantages: it's useless outside the home and the infrastructure is significantly more expensive. The infrastructure is significantly more expensive, once. The cost to maintain fiber is not materially different than the cost to maintain copper. And it's not useless outside the home when everybody else has one too. If you're at home you have WiFi. At work? WiFi. Friend's house? Coffee house? Hotel? WiFi. About the only fixed places there is no WiFi are the places there is also no cellular, like the middle of the woods. Using cellular when you're in a building with WiFi doesn't make sense. Using cellular when you're on the move does -- and that's the use case we should be preserving the scarce wireless spectrum for, by using wired connections everywhere else.

by lugg 12y ago

AnthonyMouse 12y ago

The problem is that wireless is a shared medium. If you have a certain amount of spectrum, you can transfer a finite number of bits per second using it, and that capacity is shared between all devices in range. The bandwidth available in a certain chunk of spectrum is a physical limit. There is no way to expand capacity other than by allocating more spectrum, and radio spectrum is extremely scarce. So the more users you have using cellular for internet, the less bandwidth they each get. The solution when you want to have "wireless" devices but there isn't enough wireless spectrum is to use wireless for the "last 50 meters" rather than the last mile. In other words, you have fiber to the premises and then WiFi within the building. By keeping the wireless power levels as low as possible you reduce the number of other users you're sharing the spectrum with and allow each user to have a satisfactory amount of bandwidth. Then reserve cellular wireless for users not in a building, e.g. accessing a map from your car, and have all manner of stupid bandwidth caps to discourage usage of cellular wireless whenever it isn't strictly necessary. > Is a mesh network an answer? No. Mesh networks would allow you to have more bandwidth per user by reducing wireless power levels and shrinking the geographic size of the collision domain, but that would require peers to act as repeaters between other peers and the tower. That's an epic fail. It adds latency, reduces battery life, impacts reliability if any peer's device is flaky, has security implications, etc. Wireless mesh networks ironically work much better for fixed-position devices than mobile devices, because they can use directional antennas, don't have the extreme power constraints mobile devices do, etc.

lugg 12y ago
Thanks mostly confirming what I suspected. I did think we would be able to use the spectrum more efficiently over time with better, more precise measurements but my guess is that will scale to its own limits a lot faster than a hard line will.
- martinald 12y ago
  Much like procesing power and moore's law, we're getting very close to shannon's law (http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem. Wireless isn't this huge gamechanger that people may think. To beat shannon's law you would need so many base stations in the long term it looks awfully like what we have now with DSL/Fiber/DOCSIS and wifi APs.
exelius 12y ago
It all depends how much bandwidth you need for a given use case. For most use cases other than video, existing wireless technologies are sufficient and will only continue to improve as things like beam shaping and micro cells become more prevalent. Wireless is easier to deploy and as its use grows, more spectrum will likely open up. Free markets have the benefit of deploying scarce resources where they are most in demand. Implementing that spectrum is often as easy as a configuration change that can be deployed remotely. The biggest advantage of a wired home broadband connection over a wireless one that I see is the ability to stream video. A wired connection has a lot of disadvantages: it's useless outside the home and the infrastructure is significantly more expensive. Wireless has signal and coverage issues for sure, but wireless is a much more competitive industry than wired access in no small part because it's cheaper to deploy. I don't think it would be going out on a limb to say the HN audience is much wealthier and more technology-literate than the average citizen, and thus would be more likely to want multiple Internet connections and have more use cases such as running file servers, etc. I think the HN crowd overestimates what people use their Internet connections for; the average person uses their home broadband connection primarily to watch video. It's not a bad use case by any means; but I'm not sure the government needs to be building infrastructure that will be primarily used by video streaming companies. Traditional cellular wireless (which can use any mix of wireless, wired or satellite backhaul depending on the application) and Wi-Fi are flexible enough to cover the "access" part of Internet access; especially in an increasingly post-PC world.
- AnthonyMouse 12y ago
  > For most use cases other than video, existing wireless technologies are sufficient and will only continue to improve as things like beam shaping and micro cells become more prevalent. Excluding today's most demanding use case doesn't get you anywhere. If you need a wired connection for that then you need a wired connection. The installed base of smartphones and tablets is still growing significantly and that's going to eat any technological improvements you can make to cellular and then some. Meanwhile people still want to be able to watch Netflix on a tablet from the back of a minivan, and to be able to make that happen at scale you're going to have to remove as much other traffic from the cellular network as possible. > Wireless is easier to deploy and as its use grows, more spectrum will likely open up. Free markets have the benefit of deploying scarce resources where they are most in demand. Free markets can't change the laws of physics. Allocating more spectrum is a linear increase. 20% more spectrum costs billions of dollars and only makes it 20% faster. The market solution when there is a high demand scarce resource is to shift usage to a resource that isn't as scarce for any use case where the alternative resource is viable. There is a reason one of the first things you already do when you get to a friend's house for the first time is ask for the WiFi password, and it's not because there is no cellular coverage. > The biggest advantage of a wired home broadband connection over a wireless one that I see is the ability to stream video. The advantage is that it allows you to transfer more bits cheaply. Today's most prevalent use case for that is streaming video, sure. It's hardly the only one. Anything that requires uploading or downloading large amounts of data is a bad fit for cellular. And we can't predict what interesting applications will become commonplace once the typical connection is 1Gbps. > A wired connection has a lot of disadvantages: it's useless outside the home and the infrastructure is significantly more expensive. The infrastructure is significantly more expensive, once. The cost to maintain fiber is not materially different than the cost to maintain copper. And it's not useless outside the home when everybody else has one too. If you're at home you have WiFi. At work? WiFi. Friend's house? Coffee house? Hotel? WiFi. About the only fixed places there is no WiFi are the places there is also no cellular, like the middle of the woods. Using cellular when you're in a building with WiFi doesn't make sense. Using cellular when you're on the move does -- and that's the use case we should be preserving the scarce wireless spectrum for, by using wired connections everywhere else.