Link64, on 07 July 2012 - 01:08 PM, said:
And you are apparently one more of those, who think so. NOTHING except for the first few uploads at the beginning of the eMule session of the uploader will finish sooner. So in case of your 60KB/s the first 15* or so downloaders will get the requested data sooner, after that it doesn't matter. It will start later, run at higher speed and finish at the same time, as it would if it has started earlier and run at slower speed.
*) properly configured eMule should not open more slots than (about) that at 60KB/s, I have 12-14 at 50KB/s.
I disagree to that. But of course it doesn't really matter because in most cases it will only be partial downloads. The waiting times are usually much longer than the download times anyway. If either one of us is right the difference will only be like an hour or so. That is nothing compared to the month I have to wait for some files to download.
Of course sequential download will benefit the first in line most. If you have on average 15 simultaneous downloads then #1 in line will be done the fastest where #15 will have no gain but also no loss. But after that when the next 15 downloaders get their turn the thing starts at 0. So #16 gets it sooner and #30 won't have any gain. By the time #30 starts he will have 29 chunks extra available to request from other users. So, will he need me? Or kindly leave my bandwidt to #31?
If people did not start at the same time but at a 3 minute interval at 3 kB/s things would be different. If that pace is broken and new additions are halted at time 0 till everybody is finished, people who are finishing their downloads will have that done sooner. The average total remaining chunks of the 30 requested at time 0 will be 15. If one user finishes it the other users will get his bandwidth so the 15 chunks will all be done in 30 minutes.
User #1 will be the one who in the old situation was due for upload at time 0. He would have had 3 kB/s so he would have finished at time 1:00 hour.
Now he has to wait till everybody is finished and get 60 kB/s so user #1 will start at 0:30 and end at 0:33. He will gain 27 minutes.
User # 2 would have finished at 1:03, but now starts at 0:33 and ends at 0:36. He will gain 27 minutes
User # 3 would have finished at 1:06, but now starts at 0:36 and ends at 0:39. He will gain 27 minutes
User # 4 would have finished at 1:09, but now starts at 0:39 and ends at 0:42. He will gain 27 minutes
User # 5 would have finished at 1:10, but now starts at 0:42 and ends at 0:45. He will gain 27 minutes
User # 6 would have finished at 1:15, but now starts at 0:45 and ends at 0:48. He will gain 27 minutes
User # 7 would have finished at 1:18, but now starts at 0:48 and ends at 0:51. He will gain 27 minutes
User # 8 would have finished at 1:21, but now starts at 0:51 and ends at 0:54. He will gain 27 minutes
User # 9 would have finished at 1:24, but now starts at 0:54 and ends at 0:57. He will gain 27 minutes
User # 10 would have finished at 1:27, but now starts at 0:57 and ends at 1:00. He will gain 27 minutes
User # 11 would have finished at 1:30, but now starts at 1:00 and ends at 1:03. He will gain 27 minutes
User # 12 would have finished at 1:33, but now starts at 1:03 and ends at 1:06. He will gain 27 minutes
User # 13 would have finished at 1:36, but now starts at 1:06 and ends at 1:09. He will gain 27 minutes
User # 14 would have finished at 1:39, but now starts at 1:09 and ends at 1:12. He will gain 27 minutes
User # 15 would have finished at 1:42, but now starts at 1:12 and ends at 1:15. He will gain 27 minutes
User # 16 would have finished at 1:45, but now starts at 1:15 and ends at 1:18. He will gain 27 minutes
User # 17 would have finished at 1:48, but now starts at 1:18 and ends at 1:21. He will gain 27 minutes
User # 18 would have finished at 1:51, but now starts at 1:21 and ends at 1:24. He will gain 27 minutes
User # 19 would have finished at 1:54, but now starts at 1:24 and ends at 1:27. He will gain 27 minutes
User # 20 would have finished at 1:57, but now starts at 1:27 and ends at 1:30. He will gain 27 minutes
User # 21 would have finished at 2:00, but now starts at 1:30 and ends at 1:33. He will gain 27 minutes
User # 22 would have finished at 2:03, but now starts at 1:33 and ends at 1:36. He will gain 27 minutes
User # 23 would have finished at 2:06, but now starts at 1:36 and ends at 1:39. He will gain 27 minutes
User # 24 would have finished at 2:09, but now starts at 1:39 and ends at 1:42. He will gain 27 minutes
User # 25 would have finished at 2:10, but now starts at 1:42 and ends at 1:45. He will gain 27 minutes
User # 26 would have finished at 2:15, but now starts at 1:45 and ends at 1:48. He will gain 27 minutes
User # 27 would have finished at 2:18, but now starts at 1:48 and ends at 1:51. He will gain 27 minutes
User # 28 would have finished at 2:21, but now starts at 1:51 and ends at 1:54. He will gain 27 minutes
User # 29 would have finished at 2:24, but now starts at 1:54 and ends at 1:57. He will gain 27 minutes
User # 30 would have finished at 2:27, but now starts at 1:57 and ends at 1:00. He will gain 27 minutes
I am not a mathematician so the actual gain may be 30 minutes. Not sure.
That was a lot of copy/pasting to bring home a simple point:
If you upload sequentially the downloaders have to wait only for the people in the queue before them.
If you upload simultaneously they also have to wait for the people in the queue
behind them.
Hence the extra delay of simultaneous slots.
I am new on this site so I have a limited number of posts per day at my disposal. I will have to be careful not to discuss to wildly or I will spend them all too fast.
Sorry for that