Inefficient tracker?

[renzar]

I was downloading a torrent that had one seed and 113 peers. The demonoid tracker was very slowly downloading simultaneously to all the peers increasing their percentages by about .1percent each then going through the whole list again for another .1 percent.

Wouldn't it be a lot quicker to upgrade several of them to 100 percent so they could start seeding and reduce the download time for all the rest? (Also safer because if the one seed were to go off line all downloading would have to halt.)

[rnr]

The seeder is in initial seed mode.

[DreadWingKnight]

Wouldn't it be a lot quicker to upgrade several of them to 100 percent so they could start seeding and reduce the download time for all the rest?

No. And what was going on was NOT the seed uploading to everyone the .1 percent. What's going on is that the seed is uploading small percentages, and all the downloaders are catching up to each other faster than the seed can spew pieces out.

And the tracker has no control over speeds.

[renzar]

I don't understand #2 since it lacks any details.

#3 doesn't seem like what was happening since the peers all increased by .1 percent before any of them increased beyond that, then the cycle started all over again. They were all kept at the same level. The tracker decides who to send to and could send to those with the fastest uptake speed if it chose to. This would generate more seeders and allow a faster distribution of the file.

[DreadWingKnight]

The tracker decides who to send to and could send to those with the fastest uptake speed if it chose to

Actually, it can't. It doesn't have the information required to make it happen.

There are many factors you're not looking at, and many factors that you have blatantly incorrect information about.

[renzar]

#5. How do you know how much information the tracker has? My program has the information. The tracker would be more efficient if it had the information. Maybe that's part of the problem.

[DreadWingKnight]

I know because I write tracker software. The tracker is merely an entry point into the swarm, and no the swarms wouldn't be more efficient if they had the information because of how much bandwidth they'd be wasting in keeping the information updated and the swarms co-ordinated.