Advertisement
Sensei-Hanzo

Satellite Internet

Sep 6th, 2018
194
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 11.84 KB | None | 0 0
  1.  
  2. Ikkitrix yesterday at 9:34 AM
  3. Everyone's been having connection issues
  4. But on the other hand, I'm able to play with my friend in the Philippines who has literal jungle internet relatively fine
  5.  
  6. Dan yesterday at 9:35 AM
  7. "Jungle Internet" lmao
  8.  
  9. Ikkitrix yesterday at 9:35 AM
  10. Somewhat better than expected tbh
  11.  
  12. Dan yesterday at 9:36 AM
  13. oh I don't doubt it. It's not "American Internet" at least
  14. just that expression is hillarious imo
  15.  
  16. Ikkitrix yesterday at 9:36 AM
  17. And in his new place he keeps track of power outages by the month instead of by the day
  18. Oh, it's still not great internet, but probably better than most rural places in the us, lol
  19. I really really want the low earth orbit satellite internet to work as well as it sounds
  20. And just move somewhere remote af
  21.  
  22. Dan yesterday at 9:40 AM
  23. huh not heard of "low earth orbit satellite internet", I'm assuming that's different/better than traditional satellite internet? (which doesn't even offer an upstream channel so it's still usually paired with dailup or dsl for upload)
  24.  
  25. Ikkitrix yesterday at 9:41 AM
  26. Well, let's go through the differences
  27. Current satellite internet is in geo stationary orbit
  28. Which is something like 50k kilometers away
  29. Resulting in 600ms or so in round trip latency because of the speed of light
  30. Along with being very old because launching stuff was very expensive
  31. Low earth orbit is about 2k kilometers away
  32. (These are rough ballpark figures because i don't remember exactly)
  33. But its about 20x closer
  34. But to cover all that you need to cover, you need 900+ satellites instead of 1
  35. Elon Musk has a plan that invokes launching 4000 some satellites for this purpose
  36. But theres some other companies doing less
  37.  
  38. Dan yesterday at 9:47 AM
  39. hmmm that sounded better until you mentioned Elon Musk doing it, now it sounds like an awful plan
  40.  
  41. Ikkitrix yesterday at 9:48 AM
  42. Lol
  43.  
  44. levitas yesterday at 9:48 AM
  45. As I understand it, and LEO solution would both be inherently global and also give the Pacific ocean significantly more internet capability than its had before
  46.  
  47. Ikkitrix yesterday at 9:48 AM
  48. This is true
  49. You could probably focus the constellation around certain points
  50. By how you choose the orbits
  51. But like, a theoretical 30ms round trip instead of 600ms limiting factor due to the laws of physics combined with massive redundancy and modern tech
  52. Could be amazing
  53.  
  54. Dan yesterday at 9:51 AM
  55. hmmm but doesn't the pacific ocean already have one (or more) of the strongest internet pipelines at its bottom? how're sattelites supposed to beat that?
  56.  
  57. Ikkitrix yesterday at 9:52 AM
  58. Sadly, i doubt there's many switches out in the middle of the ocean
  59. So the angler fish might have dank internet
  60.  
  61. Dan yesterday at 9:52 AM
  62. pfft. I bet the NSA has something to say about that
  63.  
  64. Ikkitrix yesterday at 9:52 AM
  65. But the surface fish are fucked
  66.  
  67. Dan yesterday at 9:53 AM
  68. FUCK YEAH ANGLER FISH INTERNET
  69.  
  70. Ikkitrix yesterday at 9:54 AM
  71. So there's this company
  72. http://www.oneweb.world/
  73. OneWeb
  74. Shouldn’t everyone have access to the world's information? In 2018 OneWeb will launch a constellation of satellites to provide affordable high-speed Internet access for the world’s unconnected
  75.  
  76. And spacex's thing
  77. Im not sure how you compete with spacex in this though, lol
  78.  
  79. levitas yesterday at 11:03 AM
  80. Worth noting that LEO doesn't get the same advantage as geostationary for covering particular locations. You could do something through the poles and focus on a particular longitude, but that's probably as good as you could get, and if you're not the only one doing it there would be a higher probability of collision at the poles (not sure if that's a real problem though given that space is big)
  81.  
  82. Ikkitrix yesterday at 11:05 AM
  83. Yeah, space is way too big to be concerned about collisions
  84. I thunk they plan on having them operate at several heights
  85. So anything that would cross the same path would actually be miles apart still
  86.  
  87. levitas yesterday at 11:07 AM
  88. Only mentioned it because there's actually convergence points for all different polar routes, whereas that problem doesn't exist at all on more standard orbits
  89.  
  90. Dan yesterday at 11:09 AM
  91. don't they still track the orbits of everything larger than a few inches up there to avoid collisions? they wouldn't do that if there weren't a possibility of problem, right?
  92.  
  93. Ikkitrix yesterday at 11:10 AM
  94. Yes, space junk is an issue, but planned orbits when it's all stuff that you're putting up there isn't really a problem
  95. Like, the problem of crashing your own satellites into each other shouldn't be an actual problem unless someone really fucks up
  96.  
  97. Dan yesterday at 11:11 AM
  98. yeah, but if there's competing networks, like Dave said "if you're not the only one doing it"
  99. then they're like space junk to each other, except with planned points of convergence
  100.  
  101. Ikkitrix yesterday at 11:13 AM
  102. You still should have many miles of separation to work with per satellite with intersecting paths
  103. And i don't think either one can do anything without some approval from some regulating body, but i don't know how this works really(edited)
  104. I think the fcc already approved the musknet
  105. But i don't know what else there is
  106. And whay happens if China feels like launching their own, etc
  107. Im not really concerned if it's just these few us based companies
  108. But space junk is a huge problem that seems way harder to fix than it is to cause and nobody really has enough incentive to address it right now
  109.  
  110. Dan yesterday at 11:21 AM
  111. true
  112. if we can't even handle junk on the planet, how're we gonna clean up the space junk? lol
  113.  
  114. Ikkitrix yesterday at 11:24 AM
  115. Especially when millimeter sized particles going orbital speeds can cause massive damage
  116. At least in ocean cleanup you don't have to worry about your craft getting destroyed by a rouge flying plastic bag
  117.  
  118. Dan yesterday at 11:26 AM
  119. but it's funnier if you do worry about that
  120.  
  121. Ikkitrix yesterday at 11:26 AM
  122. Makes it more interesting, for sure
  123. So i just remembered that those latency figures from before were sone worst case latency napkin math
  124. Because i was at my grandparents house and they have actual satellite internet
  125. Speedtest.net was giving me actual 600ms ping
  126. LEO is around 20x closer, so thats how i got to 30ms
  127.  
  128. Dan yesterday at 11:34 AM
  129. hmmm I'm not sure ping works like that though
  130.  
  131. levitas yesterday at 11:35 AM
  132. Is it going to be better than that though? I'd expect the path to be something like home > GS sat > ISP on ground > GS > home
  133. NVM I'm dumb, distance traveled would still be cut to 5% in that case too
  134.  
  135. Ikkitrix yesterday at 11:36 AM
  136. Right, so it's the same for both
  137.  
  138. Dan yesterday at 11:36 AM
  139. with current real satellite internet, only 1/2 your trip is via satellite, if LEO internet is like a glorified LTE network with the antennae in space or something, you'd have both directions going through space
  140.  
  141. levitas yesterday at 11:37 AM
  142. Right
  143.  
  144. Ikkitrix yesterday at 11:37 AM
  145. Except gso is 35000km or .117 light seconds
  146. So even if both channels went through the satellite
  147. Physics would limit round trip to 450ms
  148. Not the actual 600ms that i got
  149.  
  150. Dan yesterday at 11:38 AM
  151. ping is usually between 15-30% of the limit in my experience
  152.  
  153. Ikkitrix yesterday at 11:38 AM
  154. And leo is everything up to 2000km i guess?
  155. I don't know the actual planned orbit height
  156.  
  157. Well, ok
  158.  
  159. Dan yesterday at 11:41 AM
  160. also depends on where it's going, does it have to go sat to sat? or is it high enough to always get to a land ISP with a single sat bounce?
  161.  
  162. levitas yesterday at 11:41 AM
  163. That lag spike when you switch satellites because one is hiding in the sun's glare though lol
  164.  
  165. Ikkitrix yesterday at 11:43 AM
  166. I imagine you'll constantly be switching satellites
  167. And be in range of a couple at any given time
  168.  
  169. Dan yesterday at 11:43 AM
  170. that doesn't make it better :v
  171.  
  172. Ikkitrix yesterday at 11:44 AM
  173. It will have to go sat to sat for longer distances
  174.  
  175. Dan yesterday at 11:44 AM
  176. your latency would be all over the place lmao
  177.  
  178. Ikkitrix yesterday at 11:45 AM
  179. I mean, the switching problem is a technical issue that im not sure how much of which can be smoothed out
  180.  
  181. Dan yesterday at 11:45 AM
  182. it's probably kinda like driving all day long on your LTE network
  183.  
  184. levitas yesterday at 11:46 AM
  185. It'd be more predictable in theory, but yeah
  186. Schedule online matches to occur when you know the other team will be switching sats
  187.  
  188. Ikkitrix yesterday at 11:47 AM
  189. I think there's a couple differences
  190. I would imagine that the receivers are engineered with constantly switching in mind
  191. And for it to not suck
  192. I don't think it has to suck as long as there's no compete interruptions where no satellite is in range
  193.  
  194. Dan yesterday at 11:50 AM
  195. so are LTEs, although you could use multiple at once which phones don't usually do, and just switch to closest with no spike, but I just meant that your latency would go up and down as the closest gets closer and farther
  196.  
  197. Ikkitrix yesterday at 11:51 AM
  198. Oh ok
  199. But still
  200. My upper limit for acceptable latency for most gaming things is more around 80ms
  201. So proving to myself that leo satellite internet in theory could be well under that was interesting
  202.  
  203. Dan yesterday at 11:54 AM
  204. I mean, 100+ is still decently playable, it's just having a consistent ping that matters the most, so the timing you're expecting is the same, so even if you're slowly up and down between 30-80ms it'd probably make for an unpleasent gaming experience, I would hypothesize
  205.  
  206. Ikkitrix yesterday at 11:55 AM
  207. Hm
  208. Couldn't you easily solve that if you know the normal upper limit?
  209.  
  210. levitas yesterday at 11:56 AM
  211. Sound some bad math, latency flux could be really fine
  212. Wait, I have to do some better math
  213.  
  214. Ikkitrix yesterday at 11:56 AM
  215. also i don't think the variance would be nearly that much due to satellite switching due to orbit paths
  216. Depends on that actual orbit paths and how many satellites are in range at any given time
  217.  
  218. levitas yesterday at 11:59 AM
  219. 650 satellites evenly spread on a 1200 mile orbit would give roughly 200 miles from one sat to its nearest neighbor
  220.  
  221. levitas yesterday at 12:01 PM
  222. That's under 10 degrees deviation from the ground, but a difference in up to 1.2x the distance from the nearest satellite to the next one
  223. In practice they'll be packed closer because people don't really live /that/ far north and south, so even global coverage isn't necessary
  224.  
  225. Ikkitrix yesterday at 12:02 PM
  226. Also, even oneweb plans for 900 now
  227.  
  228. levitas yesterday at 12:03 PM
  229. Oh, 750 mile orbit, 1200 km, so it's even better
  230.  
  231. Ikkitrix yesterday at 12:03 PM
  232. And musknet is 4000+ with planned clustering around high use areas
  233. I wonder what theoretical ping to the other side of the world is(edited)
  234. Hm.
  235. A lot.
  236. So fucking off to the Philippines would probably still suck
  237. Unless you feel like playing on korean and jp servers
  238. Some remote-ass continental us would still be sick though
  239.  
  240. levitas yesterday at 12:12 PM
  241. Hahaha
  242. Live the new age hermit dream
  243.  
  244. Ikkitrix yesterday at 12:13 PM
  245. Yeeeaaah dude
  246. I need like
  247. Somewhere as remote as possible but still within 2 hours of an Amazon warehouse for Amazon now
  248. Wherever that place is
  249.  
  250. Dan yesterday at 12:14 PM
  251. oh wait, I forgot about what's probably the biggest killer of latency in a system like this, packet loss, it's the reason that wireless and even wifi is so much garbage compared to ethernet, and probably how you measured something like 10x the theoretical for satellite
  252.  
  253. Ikkitrix yesterday at 12:15 PM
  254. This is possible
  255.  
  256. levitas yesterday at 12:16 PM
  257. We could just melt away all those clouds
  258. Call it the Martian approach
  259.  
  260. Ikkitrix yesterday at 12:20 PM
  261. Just uh
  262. Live at the top of a mountain
  263. :thinkspin:
  264.  
  265. Dan yesterday at 12:22 PM
  266. and stop all the other electromagnetic emissions from earth, and stop lightning storms from occurring, and block the sun for good measure. Also figure out how to stop the cosmic rays from coming between you and the satellite
  267. ok, so basically just run a fiber from you to the satellite
  268.  
  269. Ikkitrix yesterday at 12:23 PM
  270. Sounds like a plan
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement