It’s like Azureus on steroids, claiming to increase download speeds by a startling 70% but, at what costs?
.First off I’d like to say that I’m a little hesitant to report this new suped-up version of Azureus as it has the potential to degrade download speeds for everybody.
Having said this, the news IS the news, and word of this new client is making its way across the net.
The basic premise behind BitTyrant is that it makes active decisions as to how many and with what peers to transfer data to, as compared to current BitTorrent client servers that exchange data according to a predetermined number of peers and settings.
The current data transfer protocol doesn’t factor in the strength or capacity of individual peers. It’s tit-for-tat irregardless of upload capacity.
Well, with BitTyrant, a modified version of Azureus 2.5, a “dynamic adjustment algorithm” is incorporated that “…maintains estimates of the rate at which peers will provide data…and the rate required to earn reciprocation., ” and using this data selects “…highest capacity peers and send(s) them data at the minimum rate that will cause them to reciprocate.”
The creators claim that this dynamic adjustment algorithm modification makes it much, much, faster than its regular Azureus 2.5 cousin.
During evaluation testing on more than 100 real BitTorrent swarms, BitTyrant provided an average 70% download performance increase when compared to the existing Azureus 2.5 implementation, with some downloads finishing more than three times as quickly.
It’s a pretty bold claim but, after testing it out for myself, I have to agree.
I grabbed a torrent from TorrentSpy which was 350MB in size and had 1200 seeders and 1300 leechers.
Now normally I would average about a 70 kB/s DL speed or so for a total download time of around 30min.
With BitTyrant I averaged speeds of around 300kB/s, and it finished downloading in about 12 minutes.
This is all performed with a max 875 kB/s broadband internet connection.
Now, these statistics aren’t scientific by any means but, for those of you who use public torrent tracker sites like TorrentSpy, you’ll agree that 12 minutes for a 300MB file is pretty darn fast.
What’s also interesting to point is that with Azureus you have the option of throttling the UL speeds PER/torrent tracker. By using BitTyrant you then get a maximized DL speed based on that THROTTLED UL speed.
As an example, I throttled a torrent tracker down to a 20 kB/s max UL speed. Even on a PUBLIC torrent tracker site, with 60 seeders and 113 leechers, it still managed to get an average 120 kB/s DL speed! Pretty darn amazing if I say so myself.
Is it fair to use BitTyrant? Will it harm the BitTorrent community? On the surface I’d like to say yes it does but, it does make a pretty good argument for itself.
The creators make the point that if a user is getting data from you at a rate of 30kB/s, then offers you less in return that it’s unfair to you and wastes your precious upload capacity on somebody who is not mutually beneficial.
Here’s a few screenshots of the dynamic adjustment algorithm in action.
Screenshot from the site:
It constantly reassesses the “relationship” between yourself and each of the persons you’re connected to in a torrent swarm.
Is it fair to you? Yes. Fair to others? Kind of. Fair to the? Probably not.
Some people have lousier connections than others, and so to start “hoarding” bandwidth and only sharing it with “worthy” people can have a seriously destructive effect on the file-SHARING and the BitTorrent community in particular.
From the client’s FAQs:
Q: Won’t BitTyrant hurt overall BitTorrent performance if everyone uses it?
This is a subtle question and is treated most thoroughly in the paper. The short answer is: maybe. A big difference between BitTyrant and existing BitTorrent clients is that BitTyrant can detect when additional upload contribution is unlikely to improve performance. If a client server were truly selfish, it might opt to withhold excess capacity, reducing performance for other users that would have received it. However, our current BitTyrant implementation always contributes excess capacity, even when it might not improve performance. Our goal is to improve performance, not minimize upload contribution.
So who knows what the answer is for sure but, “maybe” is certainly not a good enough reason to start screening with whom and how much you share.
If we start being overly selective we’ll begin to have our own P2P caste system, with AOL and dial-up as the outcasts and the guys with the T1 connections serving as the gatekeepers of precious data download streams.
For those that disagree, and think that it’s only fair to get an equal amount of data in return for what you upload to others, BitTyrant is easy to set up and configure.
It’s basically Azureus 2.5, and a guide on setup and installation can be found here.