TopBT on Vuze - A Topology-Aware BitTorrent Client
January 23, 09 by sharky 8,914 views
In an article published back in October, we reported about a new technology for BitTorrent called TopBT. TopBT was originally built on LH-ABC, an open-source BitTorrent client - but it has been vastly improved into a new version of Vuze. Unlike most currently existing popular BitTorrent clients such as Vuze, BitTorrent™, and µTorrent, the unique feature of TopBT is, it actively discovers its network proximity to its connected peers. It does so through actively sending packets to those peers, and inspecting their responses. Besides the topology-awareness, TopBT also considers network transmission rates to its peers, so that it can download fast. Not only that, but TopBT also saves unnecessary BitTorrent traffic that clogs up the Internet, which is great for those who are bandwidth-capped by their ISP. But is TopBT faster than µTorrent? In our tests: Yes.
There are two main goals that TopBT aims to achieve. First, it wants to reduce its generated network traffic due to topology-unawareness. This is very important for Internet traffic saving, because BitTorrent traffic accounts for a very high percentage of the total Internet traffic, more than 40% as reported in an IPOQUE report in 2007. Second, it wants to maintain fast download time, a download time that can be comparable to the existing popular BitTorrent clients on average, if not faster than them.
TopBT is implemented based on open-source Vuze code base version 4.0.0.4. Our evaluation results from typical runs on Windows XP show that TopBT can achieve comparable download speed compared to the original Vuze and uTorrent, sometimes even faster. More importantly, it can save Internet traffic significantly…
TopBT Installation
TopBT is almost identical to Vuze in every way, except there’s no "Vuze View" - only the classic (Azureus) view is available. By default TopBT is started in "Advanced" mode, which is required in order to enable and modify TopBT settings.
1. Download TopBT here. It’s available as a Windows installer, or as a .jar file for Linux. Note that on Linux and Unix, you need "root" priviledge to run the TopBT jar file.
2. TopBT requires Nmap (includes WinPcap library) installed on your computer before you can use the recommended TCP Ping based unchoking mechanism. Nmap for Windows can be found here, while other versions can be found on their homepage. TopBT also requires Java (JRE).
Enabling TopBT Settings
Open TopBT (aka ‘Vuze’) and go to Tools > Options… You’ll see a tab on the left called TopBT - click it. The default is set to OFF - choose either "TopBT Ping" or "TopBT TCPPing" (try them both out). Be sure to click SAVE at the bottom.
NOTE: TopBT comes well-equipped with some decent default client settings; not much needs to be configured. However, you may want to use the "Configuration Wizard" in the Tools menu, to set up your ISP connection settings and port testing.
ALSO NOTE: TopBT uses a speed setting called "Auto-speed (beta)", which may report your UL/DL capabilities much lower than they actually are. We found that leaving this setting to (beta) works best, although you may need to manually change the Upload Limit and Download Limit on that page. Find it in Options > Transfer > Auto-Speed.
Viewing Your "Saved" Bandwidth
With TopBT running, double-click in the bottom-right corner where it says, "TopBT saved…". This will open up the client statistics. To view details about your ’saved’ data and protocol traffic, click on the "Traffic" tab:
Speed Tests: uTorrent vs. TopBT
We tested a variety of different popular torrents from Mininova’s "Featured" torrents, and in at least 70% of the trials, TopBT took the top spot for download speed (by finishing the torrents first). Both clients were configured comparably (DHT = ON; UL limit = 100 KB/s; DL limit 500 KB/s; with similar per-torrent and global connection settings), and identical torrents were added to both clients simultaneously.
For anyone looking to increase their download speed on public trackers, and save some bandwidth in the process, TopBT is a BitTorrent client that we highly recommend. But don’t take our word for it: try it; you’ll see.
©2009 • XHTML • 
01 • REy Says: 24.01.09 at 12:46 am
Is this client allowed on pvt trackers ?
02 • Dawg Says: 24.01.09 at 1:58 am
Hey man.. I tried Topbt on a Private tracker… it gave me speed of 120 kbps.. where as utorrent gave the speed of 1mbit with the same torrent….. I guess nothing can beat Utorrent
03 • Diji1 Says: 24.01.09 at 3:02 am
It would have been preferable to run these comparisons with DHT turned off I would’ve thought. It adds another variable that can skew results. Haven’t you got the situation where, hypothetically and for example, TopBT can connect to an available DHT source which uTorrent can then no longer connect to?
It would also be good IMO to give more details of the testing methodology being used.
04 • Perhaps Says: 24.01.09 at 3:06 am
“I guess nothing can beat Utorrent :D”
Well, have a look at this test: http://filesharefreak.com/2008/03/01/which-bittorrent-client-is-the-fastest/
It only scored the slowest of all.
However, I think it must be said that testing over the internet is fraught with difficulties owing to the nature of how TCP/IP works - even if the author sharky has done his best as far as I can see.
05 • a/s/l Says: 24.01.09 at 4:05 am
well it’s gonna take all the torrent sites i frequent to add it to their whitelist before i can be bothered to try it
06 • Dawg Says: 24.01.09 at 11:32 pm
“I guess nothing can beat Utorrent :D”
Well, have a look at this test: http://filesharefreak.com/2008/03/01/which-bittorrent-client-is-the-fastest/
It only scored the slowest of all.
However, I think it must be said that testing over the internet is fraught with difficulties owing to the nature of how TCP/IP works - even if the author sharky has done his best as far as I can see. >>>>>>>>>>>>>>>>>>>>>>>>>
But mate.. those clients who achieved the max download rate are banned by most of the private tracker… Everyone doesnt use Public tracker… So yeah… For me Utorrent works the best….
07 • Perhaps Says: 25.01.09 at 2:51 am
2 out of 5 are commonly banned, this is correct.
08 • Trackback :: restricting upload speed on torrents - Page 7 25.01.09 at 10:23 am
[...] if u really wanna results read here TopBT on Vuze - A Topology-Aware BitTorrent Client | THE source for BitTorrent & P2P Tips, Trick… i hv tested this n found to be working up nice…..moreover it works better on public [...]
09 • Sharky Says: 25.01.09 at 6:20 pm
TopBT a banned client on private trackers? - Agreed. Hence why I didn’t mention private trackers. You don’t want to be using TopBT there - not a single one will ‘whitelist’ this client, and it’s up to you to know what BT client is allowed. I wouldn’t risk even one of my private trackers for a “comparison” with this. But let’s be frank - 90%(+) of BitTorrent users are using public trackers.
This technology is very new, thus the TopBT dev team will have a difficult time orchestrating an official BEP for uTorrent/BitTorrent. With uTorrent being closed-source software makes it all the more difficult for TopBT to be implemented, yet I believe it should be included at least as an option. But TopBT does / can/ may increase download speeds on public trackers - try it for yourself.
010 • ME Says: 25.01.09 at 9:45 pm
Some clients don’t allow multiple connections to the same torrent from the same IP, so if one client takes a peer first, the other client wouldn’t be able to connect to the same peer, even if that peer has open upload slots to spare. For all we know, it could be that TopBT is connecting to the local peers more quickly, and by the time uTorrent gets to attempting a connection with such peers, the peer denies any additional connections from the same IP. At that point, the only fault you can then place on uTorrent is that it isn’t prioritizing peers in the same way as TopBT does. And what if the clients are downloading from one another? That’d potentially skew results in unpredictable ways as well.
Basically, from the information given, this test you conducted isn’t scientific, and as such, should be taken with a grain of salt. Running the two clients simultaneously on an uncontrolled environment is far from an accurate method of testing client speeds.
011 • Dawg Says: 26.01.09 at 2:53 am
I agree with Me… This is so true…. Most of the clients Ban the Multiple connection on the same torrent for 1 ip…
012 • Dawg Says: 26.01.09 at 2:57 am
May be you are true sharky.. I Tested TopBT on Private tracker.. may be it does work fast on Public trackers.. But u know its more like Azureus it takes so much computer’s resources… Anyways thanks for the extra info @ Sharky….
013 • etan Says: 27.01.09 at 4:05 pm
Because sharky evaluated topbt and utorrent by running them
simultaneously on one machine, thus the comment questions that it is
not a scientific way to compare the performance, since for a nearby
(local) peer, it might only allow one connection from the same IP at a
time. This is true since the option for `bt.allow_same_ip - Allows
multiple connections from the same IP’ is default to false on utorrent
or Vuze as I checked. However, I think the chance of locating to the
same peer is small when the peer pool is about thousands peers,
although I agree it is a better way to evaluate on two machines while
also prevents them from connecting to each other by banning the IPs.
When I did the evaluation, I always used two machines, and also banning the connection between the two machines once I found any (quite small chance). –etan (developer from topbt)
014 • ME Says: 27.01.09 at 7:55 pm
Indeed, I’ll quote what I posted in the comments section for the other performance comparison post:
“That’s not to say that the experiment doesn’t provide interesting results, but it means that you can’t take away from the experiment any absolute fact about which client is fastest.”
I won’t argue that there are potential performance gains to be had from more intelligent peer selection algorithms, but the experiments attempting to demonstrate this would probably be less questionable and have a stronger impact if more wrinkles were removed
015 • Princeton Says: 01.04.09 at 6:44 pm
ME suggests that TopBT’s ability to connect to better peers faster than uTorrent may be the reason why TopBT performed better, but this supposition is reason enough to conclude the TopBT is faster the uTorrent.
If mouse A approached a piece of cheese before mouse B approached his piece, where both pieces are identical, then even if both mice eat the cheese at the same rate, mouse A will still finish before mouse B.
Now, in theory TopBT transfers data more efficiently than other clients. So in reality, to continue the metaphor, mouse A eats his piece at a faster rate than mouse B.
Not only is mouse A faster at eating his cheese; he is also faster at approaching the cheese.
The experiement was not completely unbiased because TopBT may have approached better peers faster, but this supposition alone supports the superiority of TopBT.
In other words, by supposing that the experiement was unfair for uTorrent ADMITS uTorrent’s inferiority.
New technology has surpassed the norm. This is a _good_ thing, uTorrent fanboys. For now, if you want superior speeds, use TopBT; and wait until uTorrent adopts the superior technology (such as tabbed browsing in IE.. AFAICR, Firefox had tabbed browsing first).