[freenet-support] Why is the 0.7 network slower then the 0.5?network?

[Matthew Toseland]

On Mon, Jan 29, 2007 at 07:25:48PM +, this.is.not.my.real.email.addres
s wrote:
> Matthew Toseland  writes:
> 
> > 
> > So in summary, for the small file the two networks were comparable - 
> > 1:30ish vs 2:00ish. For the ISO, I'm not sure if there was any 
> > conclusion?
> > 
> > My 2 cents: Inserts are much faster after the changes to load limiting 
> > just before xmas, or at least, they should be. Generally freenet 0.7 
> > uses close to the bandwidth limit now, which it certainly didn't before.
> > 
> > Comments?
> 
> You missed the part where that person uploaded a 500 meg file in 23 hours. 
> They also sent a subsequent message where someone else downloaded the file 
> in 18 hours.  

On 0.5? Was a similar test carried out on 0.7 though? We need a straight
comparison to draw any conclusions.
> 
> I'm running the 0.5 network now and I have been able to download some files.
> It's not as fast their downloads, but it's faster then 0.7 downloads.
-- next part --
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: 

[freenet-support] Why is the 0.7 network slower then the 0.5 network?

[Volodya]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Matthew Toseland wrote:
> So in summary, for the small file the two networks were comparable - 
> 1:30ish vs 2:00ish. For the ISO, I'm not sure if there was any 
> conclusion?
> 
> My 2 cents: Inserts are much faster after the changes to load limiting 
> just before xmas, or at least, they should be. Generally freenet 0.7 
> uses close to the bandwidth limit now, which it certainly didn't before.
> 
> Comments?

In the moments when my node is not backed off my large inserts are doing great. 
I don't
think i could achieve such speeds on the dead-net. However, those breathe 
periods between
back-off are far and few in between, so over-all, inserts are quite slow for me 
now. I
think in the past 24 hours i've managed to insert around 15-20MiB.

- Volodya

- --
http://freedom.libsyn.com/   Voice of Freedom, Radical Podcast
http://freeselfdefence.info/ Self-defence wiki
http://www.kingstonstudents.org/ Kingston University students' forum

 "None of us are free until all of us are free."~ Mihail Bakunin
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFFvIyMuWy2EFICg+0RAlhuAKDJSblHZSb1A/zr2deRDkudQ0lp7wCeLNnD
krA8xVZmXZ7r2s44oRiAFDc=
=imJP
-END PGP SIGNATURE-

[freenet-support] Why is the 0.7 network slower then the 0.5 network?

[Matthew Toseland]

So in summary, for the small file the two networks were comparable - 
1:30ish vs 2:00ish. For the ISO, I'm not sure if there was any 
conclusion?

My 2 cents: Inserts are much faster after the changes to load limiting 
just before xmas, or at least, they should be. Generally freenet 0.7 
uses close to the bandwidth limit now, which it certainly didn't before.

Comments?

On Wed, Jan 24, 2007 at 06:12:31AM -0800, this.is.not.my.real.email.address
 wrote:
> I saw the following message on Frost's Freenet board.  Scroll down to the
> end to see the results.  With my experiance uploading and downloading files
> on the the 0.7 network, there is no way I have found to make my speed even
> come close to what they can do on the 0.5 network.  Why is the 0.7 network
> so much slower?
> 
> 
> - Freenet Is Dead at Tb3nN9rhFbV6C99cN+O1fZM79rQ - 2007.01.20 -
> 04:59:37GMT -
> 
> It looks like development has stopped.
> More waste time do nothing prove nothing simulations are in the works.
> LOL!!!
> Why? Look what the sims have done for us. Nothing.
> It's beginning to look like 0.7 is doomed.
> Toad is all but silent.
> 
> - Anonymous - 2007.01.20 - 15:03:09GMT -
> 
> The 0.7 freenet is working the way Toad wanted. He didn't want open net and
> didn't want it to be used for file trading, so as far as he is concerned its
> working as designed. There is no need to do any more development.
> 
> - Anonymous - 2007.01.20 - 21:32:37GMT -
> 
> Freenet 0.7 is brilliant for file trading though!
> 
> check out all the stuff available via Thaw and advertised in the music
> boards.
> 
> - Anonymous - 2007.01.21 - 16:44:34GMT -
> 
> Brilliant for file trading? I have an outstanding challenge to any of the
> 0.7 users who would want to race. So far there have been no takers. How
> about you?
> 
> - Anonymous - 2007.01.22 - 10:20:15GMT -
> 
> OK, sure. How should we do it to ensure the most accurate results?
> 
> 1. We should test both inserts and retrieves.
> 2. We should test a range of filesizes.
> 3. We should do it several times with different files of almost identical
> sizes, so we can see the average time and how much the variance is.
> 4. We could do it with long-running nodes and freshly installed nodes (I
> understand 0.5 needs some time to warm up)
> 5. We could do it with different max memory configurations.
> 
> My internet usage is capped to a certain amount per month, so I throttle my
> Freenet bandwidth accordingly. Can you throttle Freenet 0.5 bandwidth?
> 
> We could start with a simple test. The Linux version of Firefox 2 is
> 9.2MBfrom here:
> http://www.mozilla.com/en-US/firefox/
> 
> I will unthrottle my bandwidth and insert that into Freenet 0.7 and see how
> long it takes. There is a problem in that Freenet doesn't record how long an
> insert takes, so I'll have to watch it regularly. But it will give us an
> idea. I will record the % at regular intervals. I don't have any other
> inserts or retrievals going. [See below for the results].
> 
> My setup: 1.4GHz CPU, 512MB RAM, 128MB allocated to Freenet. Linux
> 2.6.18kernel, Sun Java
> 1.5.0_10-b03. Freenet runs on a cryptsetup encrypted partition. Consumer IDE
> hard disk. No other applications running so swap isn't being used. Internet
> connection is a consumer ADSL, download speed about 4Mbps, upload about
> 500kbps.
> 
> Node Version Information
> 
>* Freenet 0.7 Build #1011 r11589
>* Freenet-ext Build #9 r11062
> 
> Node status overview
> 
>* bwlimitDelayTime: 608ms
>* nodeAveragePingTime: 221ms
>* networkSizeEstimateSession: 401 nodes
>* nodeUptime: 11h7m
>* routingMissDistance: 0.1352
>* backedOffPercent: 79.8%
>* pInstantReject: 0.0%
> 
> Current activity
> 
>* Inserts: 29
>* Requests: 99
>* Transferring Requests: 17
>* ARK Fetch Requests: 6
>* Total Output: 480 MiB (12.2 KiBps)
>* Payload Output: 335 MiB (8.58 KiBps) (69%)
>* Total Input: 432 MiB (11.0 KiBps)
>* Output Rate: 13.8 KiBps (of 100 KiBps)
>* Input Rate: 12.2 KiBps (of 100 KiBps)
> 
> Peer statistics
> 
>* Connected: 3
>* Backed off: 5
>* Disconnected: 6
> 
> Inserting firefox-2.0.0.1.tar.gz (9.2MB), no compression, default priority
> (high).
> 
> I've set the bandwidth limits to 100KBps which is essentially uncapped.
> My CPU usage was minimal (< 5%) and load average was very small (< 0.2).
> 
> Time  %Completed
>   -
> 8:05   0.0%
> 8:11   3.6%
> 8:15   6.5%
> 8:20  11.2%
> 8:25  16.4%
> 8:30  21.4%
> 8:37  27.5%
> 8:40  29.7%
> 8:45  34.5%
> 8:50  38.6%
> 8:55  42.8%
> 9:00  47.1%
> 9:05  51.6%
> 9:10  55.9%
> 9:15  60.4%
> 9:20  65.4%
> 9:25  69.5%
> 9:30  73.6%
> 9:35  78.6%
> 9:41  84.4%
> 9:45  88.0%
> 9:50  93.2%
> 9:55  97.2%
> 9:57 100.0%
> 
> So that is about 5MB per hour for insertions.
> About 100MB a day.
> About 1.4kB per second.
> About 11 kbps.
> Enough to upload an album or two a day; a 700MB DivX would take about a
>