Yeah its all not too good, and hardly able to post replies. But i tried and i am getting:
344 Kb on the main download mirror.
2.28 Mb on ut-files.com secondary main mirror.
1.40 Mb on the secondary last resort mirror.
So the result can differ alot it seems. And i never tested it before. But even the main download mirror which is the slowest(seems now) you should be able to pull in the DecayedS package (610Mb) within 30 minutes. Which is not bad seeing the fact its far over half a gigabyte. And preferably keep using the the mirror as given to prevent too much stress on the servers.
And so far the server is doing fine, and no problems are reported. But if needed simply switch to the secondary mirrors, thats of course why they are there. But when possible stick to the order they are given to spare on the servers. But as said, if a certain mirror obviously isnt working for you try switch to the next on the list.
And the number you give (72kbps) is too slow. And either you have a bad connection to that server or you caught it in a bad day. And often its helpfull to reboot your computer and try again as this solves the problem often as well. But if it persists, try the next mirror in the list untill you have a decent downloadspeed, or have patience
Is this the server download limit?
Actually i have no clue, i dont even know where i could get that info for my sites. And i dont think its mentioned anywhere. But as far as i know, my Celtic Warriors site has a monster of a speed(i know someone pulled 3600kb from it once) compared to Unreal Texture which already has a very, very, very decent speed.
The main mirror i have no clue, but is from outmost importance to keep the total bandwidth under control ! So it should be used first wherever possible. But i dont know if its a shared or a VPS server, Unreal Texture itself is shared while i know ut-files.com is a VPS server which is the best you can get but 10-20 times more expensive than shared.]
But 72 is way too slow, and incorrect, you should be able to pull out alot more even on the slowest oné.
.
.
.