3 ms·

>It's almost to the point of overtaking wired services in both data rates & cost (just get those monthly data caps up and I'll switch). The data caps can't be

>It's almost to the point of overtaking wired services in both data rates & cost (just get those monthly data caps up and I'll switch). The data caps can't be removed. Wireless has much lower bandwidth limits overall--the spectrum would be saturated if everyone decided to move to wireless. It can't happen with current technology. Not to mention latency is much higher.

cptskippy 9y ago
> Not to mention latency is much higher. One of the primary focuses of 5G is to reduce latency and LTE has reduced latency tremendously. My personal experience looking over test logs on my phone is that LTE averages ~40ms with a low of 25ms and a high of 80ms. My home internet averages ~16ms with a low of 12ms and a high of 44ms. Gaming is one of the few things where you can really tell the differences between sub 100ms latencies and so for most people wireless internet is more than sufficient.
- learc83 9y ago
  Those averages don't really tell you much. What matters in terms of quality is the consistency of the latency and packet loss. But even without latency issues, we don't have the technology to replace more than a tiny fraction of total wired bandwidth with wireless.
cesarb 9y ago
Data caps are equivalent to an horribly designed token bucket, where the tokens are all added at once at the beginning of each month. Using a proper token bucket policer/shaper should be much better at preventing spectrum saturation.
- learc83 9y ago
  Still won't be enough to move everyone to wireless. We won't be able to replace more than a small fraction of our wired bandwidth with wireless anytime in the foreseeable future.
  - ninkendo 9y ago
    I think your statement is true only if you consider the area covered by a single tower to be fixed. If a given volume of space can only handle so much bandwidth over the EM spectrum, it seems the logical solution would be keep the range smaller and put antennas (with wired backhaul) in more places. Similar to how wireless companies do a lot of special capacity work in large stadiums, etc. [1] Of course, as the range decreases and the number of towers increase, the cost dynamics start to resemble the same last-mile cost problem more and more, but at least: - you don't have right-of-way problems (presumably any company can lease EM spectrum) - You can choose the areas install more shorter-range towers based on demand [1] http://www.steelintheair.com/Blog/2017/01/immaculate_cellular_reception_big_game.html http://www.steelintheair.com/Blog/2017/01/immaculate_cellula...
    - learc83 9y ago
      What you're describing would increase bandwidth. It's basically the plan for 5g. However, it still won't come close to replacing wired bandwidth. There is a limit to how low you can reduce transmission power and still get blanket coverage therefore there is a limit to how close together these towers can be without interfering with each other, and a corresponding limitation on bandwidth for a given area. You'd need to move the towers down to something close to the range of Wi-Fi to provide enough bandwidth to come close to working. And at that range you basically have all of the exact same problems of wired internet. >you don't have right-of-way problems (presumably any company can lease EM spectrum) You still have significant right-of-way problems because you still need to provide the wired backhaul for these towers. Unless you want to piggyback on the existing wired infrastructure--in which case you need to pass laws to force existing providers to let you use their pipes. Wireless internet isn't a solution to the problems we face with ISPs. It might help to provide some limited competition, but the physical limitations prevent it from replacing wired internet.
      - cptskippy 9y ago
        > You'd need to move the towers down to something close to the range of Wi-Fi It's my understanding that LTE and 5G are designed to scale such that the density of cell towers matches the density of the population. Thus in rural areas you have fewer high powered tower and in urban centers you would have many low powered micro towers. > You still have significant right-of-way problems because you still need to provide the wired backhaul for these towers. In downtown Atlanta they have ground level LTE installations with wireless backhauls at heavily traffic intersections attached to the street lights. This is in front of the Apparel Mart, you can see the antenna arrays next to the red light and further up at the streetlight: https://www.google.com/maps/@33.7610012,-84.3876772,3a,30.8y,172.71h,107.88t/data=!3m6!1e1!3m4!1sgSGRbAGROSVBtWo_RQDclQ!2e0!7i13312!8i6656 https://www.google.com/maps/@33.7610012,-84.3876772,3a,30.8y... If you zoom in you can make out LTE on lower Array.

cptskippy 9y ago

> Not to mention latency is much higher. One of the primary focuses of 5G is to reduce latency and LTE has reduced latency tremendously. My personal experience looking over test logs on my phone is that LTE averages ~40ms with a low of 25ms and a high of 80ms. My home internet averages ~16ms with a low of 12ms and a high of 44ms. Gaming is one of the few things where you can really tell the differences between sub 100ms latencies and so for most people wireless internet is more than sufficient.

learc83 9y ago
Those averages don't really tell you much. What matters in terms of quality is the consistency of the latency and packet loss. But even without latency issues, we don't have the technology to replace more than a tiny fraction of total wired bandwidth with wireless.

cesarb 9y ago

Data caps are equivalent to an horribly designed token bucket, where the tokens are all added at once at the beginning of each month. Using a proper token bucket policer/shaper should be much better at preventing spectrum saturation.

learc83 9y ago
Still won't be enough to move everyone to wireless. We won't be able to replace more than a small fraction of our wired bandwidth with wireless anytime in the foreseeable future.
- ninkendo 9y ago
  I think your statement is true only if you consider the area covered by a single tower to be fixed. If a given volume of space can only handle so much bandwidth over the EM spectrum, it seems the logical solution would be keep the range smaller and put antennas (with wired backhaul) in more places. Similar to how wireless companies do a lot of special capacity work in large stadiums, etc. [1] Of course, as the range decreases and the number of towers increase, the cost dynamics start to resemble the same last-mile cost problem more and more, but at least: - you don't have right-of-way problems (presumably any company can lease EM spectrum) - You can choose the areas install more shorter-range towers based on demand [1] http://www.steelintheair.com/Blog/2017/01/immaculate_cellular_reception_big_game.html http://www.steelintheair.com/Blog/2017/01/immaculate_cellula...
  - learc83 9y ago
    What you're describing would increase bandwidth. It's basically the plan for 5g. However, it still won't come close to replacing wired bandwidth. There is a limit to how low you can reduce transmission power and still get blanket coverage therefore there is a limit to how close together these towers can be without interfering with each other, and a corresponding limitation on bandwidth for a given area. You'd need to move the towers down to something close to the range of Wi-Fi to provide enough bandwidth to come close to working. And at that range you basically have all of the exact same problems of wired internet. >you don't have right-of-way problems (presumably any company can lease EM spectrum) You still have significant right-of-way problems because you still need to provide the wired backhaul for these towers. Unless you want to piggyback on the existing wired infrastructure--in which case you need to pass laws to force existing providers to let you use their pipes. Wireless internet isn't a solution to the problems we face with ISPs. It might help to provide some limited competition, but the physical limitations prevent it from replacing wired internet.
    - cptskippy 9y ago
      > You'd need to move the towers down to something close to the range of Wi-Fi It's my understanding that LTE and 5G are designed to scale such that the density of cell towers matches the density of the population. Thus in rural areas you have fewer high powered tower and in urban centers you would have many low powered micro towers. > You still have significant right-of-way problems because you still need to provide the wired backhaul for these towers. In downtown Atlanta they have ground level LTE installations with wireless backhauls at heavily traffic intersections attached to the street lights. This is in front of the Apparel Mart, you can see the antenna arrays next to the red light and further up at the streetlight: https://www.google.com/maps/@33.7610012,-84.3876772,3a,30.8y,172.71h,107.88t/data=!3m6!1e1!3m4!1sgSGRbAGROSVBtWo_RQDclQ!2e0!7i13312!8i6656 https://www.google.com/maps/@33.7610012,-84.3876772,3a,30.8y... If you zoom in you can make out LTE on lower Array.