BitMate: A BitTorrent Client for Poor Bandwidth People

[Cabalo]

BitTorrent is an excellent tool for sharing large files online, which is why millions of people use it every day. In developing third world countries, however, BitTorrent usage falls far behind, mainly because the transfer speeds are not that great on low bandwidth connections. Thanks to a new BitTorrent client, funded by a grant from the U.S. State Department, this situation can look forward to positive change.

A few days ago a new BitTorrent client surfaced under the promising name BitMate. The client is developed by a group of researchers from several well respected universities who have collaborated to improve the lives of BitTorrent aficionados in developing countries.

The aim of BitMate is to drastically improve the download speeds of peers on low-bandwidth connection (5 to 20 KB/sec), to make BitTorrent more effective in places where people might need it the most. If we believe the claims of the researchers, they have succeeded in making a difference.

TorrentFreak contacted Dr.Umar Saif, Associate Professor at MIT’s Computer Science and Artificial Intelligence Laboratory and leader of the initiative, to learn more about the new BitTorrent client.

“We have spent close to 2 years experimenting with various tweaks in BitTorrent, using both real-world and synthetic swarms. BitMate is our first public release and is an ongoing project,” Saif said.

During the latest tests the researchers found that compared to traditional clients, the download speeds on low bandwith connections can receive up to a 70% boost with BitMate, while upload contributions may improve by up to 1000%.

“In our target conditions, Bitmate can almost double the download performance. At the same time, it performs at least as well as the traditional BitTorrent clients for high-bandwidth peers,” Saif noted.

The beauty of it all, is that other peers are not negatively affected by these improvements.

“BitMate enhances the performance of low-bandwidth nodes without cheating, circumventing the fairness policy of BitTorrent or adversely affecting the performance of other peers,” Saif told TorrentFreak.

Among other things, BitMate can establish this advantage by prioritizing connections to other slow peers, by minimizing cross-ISP traffic and by avoiding redundant downloads. Combined with several other optimizations, the Vuze-based BitMate client is able to speed up downloads on slow connections.

“Instead of wasting optimistic unchokes on high bandwidth peers, a BitMate client optimistically unchokes those peers that have a similar low-bandwidth. As a result, a BitMate client invests its scarce upload bandwidth on peers that are most likely to reciprocate.”

“At the same time, BiTMate leaves the tit-for-tat reciprocal unchoke policy untouched to uphold the fairness of BitTorrent. This leads to both increased performance and fairness since low-bandwidth clients can quickly form mutually beneficial peer-to-peer connections,” Saif said.

A win-win situation for all BitTorrent users, generously funded by the U.S. State Department. It’s almost too good to be true.

BitMate’s latest version was released to the public three days ago and can be downloaded for free. Although the project is aimed at developing countries, there are plenty of people in other parts of the world that are on a slow connection, and might benefit from BitMate.

Looks like a quite interesting idea and conceptually it is correct. Besides, being Vuze based is a enough reason to know it is a well coded client.

[Vorx]

sounds like a cheat client...

[anon]

Among other things, BitMate can establish this advantage by prioritizing connections to other slow peers, by minimizing cross-ISP traffic and by avoiding redundant downloads.

Reminded me of the Ono plugin for Azureus, which has existed for a while already. (I tried that, but didn't like how it continuously ran pings and traceroutes)

sounds like a cheat client...

Evidently, you didn't even read the damn article.

[KFlint]

sounds like a cheat client...

Somehow, every post from you that I read is slightly to grossly retarded...

[shipwreck]

From my understanding of the article, it basically keeps the low bandwidth users amongst themselves (by prioritising the traffic between them) while not hurting / slowing down the fast connections. Sounds like a good plan.

This could probably be implemented by the established clients, too. Since it's open source, it shouldn't be too difficult.

[TONiC]

I think we're overlooking the fact that if you have a shit connection [say 5kB/s], then you have a shit connection [you aren't going to get more than 5kB/s].

[shipwreck]

That's a given. The point is to utilise those shitty 5KB/s more efficiently than bittorrent clients do today.

[anon]

Somehow, every post from you that I read is slightly to grossly retarded...

Was that worth your 5000th post?

That's a given. The point is to utilise those shitty 5KB/s more efficiently than bittorrent clients do today.

That's correct.

[TONiC]

That's a given. The point is to utilise those shitty 5KB/s more efficiently than bittorrent clients do today.

Last I checked BT:
Shares the rarest piece first
Send more requests to those with best upload speed

Thus making it efficient anyway, also demonstrated by Facebook's use of BT to update servers. The largest waste is the connection overhead between peers, especially with uTP or whatever that BW management shite was called; hence if you want a more efficient connection, just gotta connect to less people [Funny paradox no?].

[anon]

Last I checked BT:
Shares the rarest piece first
Send more requests to those with best upload speed

Thus making it efficient anyway

The specific mechanics this works under can be found here. BT is efficient by design, but BitMate is even more so, based on the assumption you're using a low-speed connection. It aims at conserving bandwidth and choosing peers that are as little hops away as possible, and therefore have lower latencies. Sounds pretty sensible to me.

if you want a more efficient connection, just gotta connect to less people [Funny paradox no?].

That's also correct. Extra connections also bring more overhead (something BM attempts to minimize), and that's less "real" traffic available for transfers. If you can count your download speed with your fingers, it may be wise to use a similar figure for the maximum connections amount.