Local Peer Discovery: Slow downloading on Vuze as compared to utorrent

[techieey]

I had my android mobile as a seeder & Vuze & utorrent as leachers (in the same LAN), but only one at a time.

In utorrent, I got a maximum downloading speed of 2 MBps (& average 1.5 MBps) whereas I couldn't cross a speed of 250 KBps in Vuze on my laptop. I am raising a point here that Vuze should have yielded more speed than utorrent, but utorrent gave more performance despite being of different company. So Vuze's performance needs to be improved in "downloading through local peers".

I used vuze 1.8.1 on Paranoid Android, based on Lollipop 5.1; utorrent 3.4.3 (build 40298) & Vuze 5.6.1.2 on Windows 7 SP1 64-bit.

Extra Information: My ISP speed is 64 KBps. & the file I used for the above process was a torrent folder of 749 MB with 6 files & one file was of 748 MB.

[techieey]

This time I changed the setup a little bit, but the results were same: utorrent was substantially faster than Vuze. 

The setup was:
Used two different computers in the same LAN, but didn't used my mobile this time. Used the same torrent. This time I sat closer to my router.

On my desktop PC, used utorrent as a seeder, and on laptop, used utorrent as a leacher. Now the downloading speed was about 10.8 MB/s. The good thing was that downloading speed was consistent. I checked this consistency in "Networking" tab under Windows Task Manager.

Now I used same desktop PC: utorrent as a seeder but used Vuze on laptop as a leacher. Now the maximum downloading speed it touched was 5.8 MB/s, but average or stable speed was 3.1 MB/s. To check it: I opened Windows task Manager & graph for networking speed was a "heart-beat like wave" pattern.

So it is clear that vuze needs to be improved in downloading by "local peers".
The results would've been more clearer to you if I had taken a screenshot.

[GaryE]

Try it the other way around.  LOL

Try seeding with Vuze and leeching with Vuze.  Then try seeding with Vuze and leeching with uTorrent.

I bet the speeds will be the same.

uTorrent has a long history of putting stuff into their code to be biased towards other uTorrent users and to upload faster to uTorrent users than it does to users who choose other clients.  They do this in part to try to gain market share.  They have been caught several times.  You did not say which version of uTorrent you are using . . . it is possible they are up to it again and/or you are using an older version.

I have never heard of Vuze doing that . . . again I would bet the speeds will be the same or much much closer.

[techieey]

Try seeding with Vuze and leeching with Vuze.  Then try seeding with Vuze and leeching with uTorrent.
I bet the speeds will be the same.

No, speeds were not the same. Vuze to Vuze was faster & Vuze to utorrent was slower.

Vuze as seeder & Vuze as leecher           : Highest speed it touched: 5.8 MB/s & average speed:  3.5 MB/s.
Vuze as seeder & utorrent as leecher      : Highest speed it touched: 3.4 MB/s & average speed:  2.5 MB/s

To put speed performance into perspective (from my second post on this thread):

utorrent as seeder & utorrent as leecher: Highest speed it touched: 11 MB/s & average speed: 10.8 MB/s
utorrent as seeder & Vuze as leecher      : Highest speed it touched: 5.8 MB/s & average speed:  3.1 MB/s

You did not say which version of uTorrent you are using

I mentioned, see the 1st post on this thread: utorrent 3.4.3 (build 40298) & Vuze 5.6.1.2 on Windows 7 SP1 64-bit.

again I would bet the speeds will be the same or much much closer.

If you are talking of average speeds, then your argument can still be accepted: 3.5 MB/s vs 2.5 MB/s; but if you are talking about maximum or instantaneous speed, then it would be wholly wrong: Vuze to utorrent didn't crossed 3.4MB/s even once, whereas Vuze to Vuze reached 5.9 MB/s & was almost varying about 4 MB/s.

I've given enough data, now other users on this forum do these speed tests on their own level & post results, to decide if some improvement needs to be done in Local Peer downloading.

[techieey]

utorrent as seeder & utorrent as leecher: Highest speed it touched: 11 MB/s & average speed: 10.8 MB/s
 

I again tested for the above setup: this time: Highest speed it touched: 9 MB/s & average speed: 7.0 MB/s. Speed is comparitively less from the previous session, but still much higher than Vuze to Vuze: average speed: 3.5 MB/s.

Some screenshots:

Vuze as seeder & Vuze as leecher           : Highest speed it touched: 5.8 MB/s & average speed:  3.5 MB/s.
 

Vuze to Vuze

Vuze as seeder & utorrent as leecher      : Highest speed it touched: 3.4 MB/s & average speed:  2.5 MB/s
 

Vuze to utorrent
utorrent as seeder & utorrent as leecher: Highest speed it touched: 9 MB/s & average speed: 7.0 MB/s
utorrent to utorrent

[GaryE]

Oops I missed the version number in the first post!  My bad!!!

As for me . . . I do not have a windows machine in my possession.  Nor do I presently have the ability to easily hook up two computers wired to my router (one is a laptop and I do not have a free wired connection right now).  Even if I had all that . . . since uTorrent for Mac uses completely different source code than does uTorrent for windows any results I got would be meaningless.

You are seeing variations in the 30% + range when testing the same setups (uTorrent to uTorrent).  10.8 MByte/sec vs 7 MByte/sec average speeds is a + 35.1% variance (7 MByte/sec vs 10.8 MByte/sec is a - 54% variance).  That is a huge variance.  If you were testing in a purely controlled environment that is the kind of variation that completely invalidates tests.

The first thing you need to figure out is the reason for that variance, then eliminate that variance, then retest!  If you can not eliminate the variance the run the test at least 15 - 20 times each and get an average and then compare the averages.

Does your ISP also provide your cable TV?

Do you have any DVR's?

Are the DVR's recording during some of the tests and not others?  Or are the DVR/Cable boxes turned on during some tests and turned off during other tests?  That could significantly alter the results.  At least with my setup the DVR's share LAN bandwidth with my computers.  I would completely unplug all DVR's and/or cable boxes to make sure they are not using any of the LAN bandwidth.

Or is your wireless network active and some wireless device is using up your bandwidth during these tests?

Are the computers in question wired or wireless?

If wireless there can be significant variation in wireless performance depending on environmental factors.  Wireless is notorious for that!  Some versions of the wireless protocol are more prone to that than others but all are effected by environmental factors.  To have a valid test the computers really should be wired.  Or if you really wanted to do the testing wireless the whole test environment would need to be contained in a Faraday cage.

Are both computers attached to the LAN using the same category of wiring?  For example is one using 10BaseT while another is using 100BaseT or 1000BaseT or whatever?

You are presenting information as fact . . . but you have not given us enough information about your testing methodology for us to evaluate your conclusions.  There are a lot of variables here.

Is absolutely everything shutdown on both computers when you are running the test?

For example if you have a virus checker installed on the computer . . . which has auto-updates enabled . . . even if the anti-virus is not running it could have been checking for new virus definitions to download and/or actually downloading updates.  There are 10's of thousands of other programs which could be running silently in the background to check for updates (not the least of which are the torrent clients themselves).  You would have to either uninstall all of the programs on both computers or better yet start from a fresh install of windows on both computers and only have uTorrent and Vuze installed.

Is your windows firewall active?  Are there exceptions in it for uTorrent and Vuze?  If there are exceptions for one or the other it could help to explain the variances.

I am not saying your conclusions are necessarily invalid . . . but I am saying that without addressing all of the variables the results are not as conclusive as you seem to think they are.  Maybe Vuze does have an issue with local peer discovery/local peer transfers but with the variance you showed  between two identical tests sorry mate I am not seeing it as a certainty.

[techieey]

Thanks for your detailed reply. I agree with all the variances that you have detailed, that may occur, or should be kept in mind. Even the small things can make a big difference, like anti-virus consuming bandwidth, etc. I agree that this slow speed issue maybe contributed to my setup only, or other people having setup similar to my setup, or even lot more people out there. So trying to conclude that this slow speed issue may be limited to me or may be experienced by a lot more people.

I'll do more tests, when I get some free time.

Does your ISP also provide your cable TV?

No

Do you have any DVR's?

No

Or is your wireless network active and some wireless device is using up your bandwidth during these tests?

Wireless network active & other devices' wifi turned off.

Are the computers in question wired or wireless?

Both. A desktop computer with a wired connection & a laptop with wireless connection.

Are both computers attached to the LAN using the same category of wiring?  For example is one using 10BaseT while another is using 100BaseT or 1000BaseT or whatever?

I don't have much information on these: I've cables that were bundled with my router & modem. Both router & PC's LAN interface are having theoretical speed of 100 Mbps.

Is absolutely everything shutdown on both computers when you are running the test?

No. Anti-virus was active, chrome notifications were also active. I exited some apps like "Bit Torrent Sync".

You would have to either uninstall all of the programs on both computers or better yet start from a fresh install of windows on both computers and only have uTorrent and Vuze installed.

I understand that this type of setup is necessary to term the results as a universal result. But I wouldn't be able to do this type of setup.

[GaryE]

Since one of your computers is using a wireless connection there are quite frankly no conclusions which can be drawn from you experiment.  As I said wireless speed depends on environmental conditions.  There are hundreds of things which can effect wireless speeds and they can change from minute to minute.  In my house I have get wireless speeds that are all over the place even when my wireless Internet device is in the same place -- that is why in the near future I will be putting a new wired connection into the location where my wife primarily uses her laptop.  I just upgraded my router hoping that the wireless issues would go away (my old wireless router was about 7 years old).  The wireless issues got better but they are still inconsistent.  Heck even the wireless speeds on my iPad go all over the place even when I am in the exact same position.

All of the rest of the stuff I mentioned you have eliminated as possible problems.  You could have an intermittent cable . . . but that is extremely unlikely!!  Since you are using commercially produced cables the odds of that being an issue are very very very . . . very very low LOL

Again I am not saying that there are no issues with local peer discovery with Vuze . . . I am just saying that the test you have run is inconclusive.  If I had the ability to repeat your test in a more controlled environment I would . . . I just do not have windows machines in my house to try and I do not have that new wire run done yet (it will not get done until my injured back feels a hell of a lot better LOL).

I do not know maybe the Vuze staff has a good test setup and they can try repeating your test.