Updated Bandwidth tester for use

Driver	Pts	±Ldr	±Pos
Cordell Cahill	52	0	0
AJ Weber	50	-2	-2
Stephan Moons	26	-26	-24
Boris Palov	20	-32	-6
Mike Redman	14	-38	-6
Daniel Paskalev	11	-41	-3
Zack Litchfield	10	-42	-1
Matt Steele	4	-48	-6
Sammy Thiels	3	-49	-1
Bill Blalock	1	-51	-2

Driver	Pts	±Ldr	±Pos
Cordell Cahill	54	0	0
AJ Weber	54	0	0
Alex du Plessis	33	-21	-21
Stephan Moons	25	-29	-8
Zack Litchfield	22	-32	-3
Mike Redman	16	-38	-6
Todd Thorne	10	-44	-6
Boris Palov	8	-46	-2
Doug Baggett	6	-48	-2
Alan Dean	4	-50	-2

Driver	Pts	±Ldr	±Pos
Cordell Cahill	177	0	0
AJ Weber	167	-10	-10
Stephan Moons	94	-83	-73
Boris Palov	89	-88	-5
Mike Redman	78	-99	-11
Daniel Paskalev	71	-106	-7
Alex du Plessis	68	-109	-3
Zack Litchfield	63	-114	-5
Alan Dean	46	-131	-17
Todd Thorne	27	-150	-19

«	February 2012	»
			1	2	3	4
5	6	7	8	9	10	11
12	13	14	15	16	17	18
19	20	21	22	23	24	25
26	27	28	29

Posted by Doug Baggett at September 02. 2010

For those of you interested....I've set up an updated version of the web100 NDT bandwidth tester here at my work location (Office of CyberInfrastructure/National Science Foundation). This software was funded by NSF and DOE and continues developement under Internet2 which is a consortium of Universities, government agencies, etc that fund that research network.

http://web100.cise-nsf.gov:7123/

Note, that if you have an ISP which does initial burst speeds the results will ONLY show what kind of speed you have as a result of the 10 second test. The above test uses java so you must have that installed.

For the more nerd/interested...

http://web100.cise-nsf.gov:8000/

This test allows you to specify a target bitrate along with latency and if you don't trust the java client even provides a link to a .c source code file you can compile and run directly against the server. The results can give you information on how to tune your host network performance based on the target latency and bitrate you desire to achieve. (emphasis..can help..not guaranteed). The software was designed with the goal of troubleshooting end to end performance issues.

I'd recommend this testing sever over speedtest.net as it's been rumored that the ISP's have been prioritizing packets bound for speedtest.net sites to make their numbers look good for their users trying to test their speeds. Also, this server is geographically close to Hoss in Maryland so even though not guaranteed due to the network path it might take chances might be good that it's an accurate reflection of the latency you might encounter to the game server. (I'm hoping)

Re: Updated Bandwidth tester for use

Posted by C at September 02. 2010

My home cable connection, 2 test avg for both

SpeedTest

down: 19.975 mb/s

up: 4.745 mb/s

Hard earned tax dollars test

down: 20.915 mb/s

up: 4.325 mb/s

That's comcast cable 15mb down 3mb up service, if I find myself downstairs I'll check verizon dsl connection.

Re: Updated Bandwidth tester for use

Posted by Scott Farm at September 02. 2010

I don't see the "start" button.

Nevermind........I had Java disabled LOL

** Starting test 1 of 1 **
Connected to: web100.cise-nsf.gov -- Using IPv4 address
Another client is currently being served, your test will begin within 45 seconds
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 1.59Mb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . 7.05Mb/s
The slowest link in the end-to-end path is a 10 Mbps Ethernet subnet
Information: Other network traffic is congesting the link
[S2C]: Packet queuing detected

Not bad for a left coaster. I pay for 8 and 1.5.

Re: Updated Bandwidth tester for use

Posted by Doug Baggett at September 03. 2010

Ignore...double post...

Re: Updated Bandwidth tester for use

Posted by Doug Baggett at September 03. 2010

I can't emphasis more that you should NOT take your bandwidth stats (from any testing site) as fact unless you do more than a 10 second test (preferably 120 seconds). For example, comcast gives me 5Mb/s upstream bandwidth (maximum during non-congested times) for 30 seconds, afterwards it drops to roughly 2 Mb/s.

Another service running on my workstation is a utility called iperf. It's a command line facility for doing more detailed testing.

The server/client can be installed using the standard apt-get/yum/portage/etc repositories in Linux and for windows at

http://www.noc.ucf.edu/Tools/Iperf/

Here is my connection doing an upstream test from my computer at home to my workstation.

-c is for client followed by the destination server

-i is for time interval of reporting (report every one second)

-t is for total time of test

So I'm connecting to jedi-rh.cise-nsf.gov, reporting every 1 second and for a total of 120 seconds

What you can see below is that after 30 seconds of roughly 5 Mb/s my upstream drops to around 2Mb/s. For rfactor assuming nothing else is running on my network I set my upstream bandwidth to half (1 Mb/s) as that is a safe bandwidth limit that has a high likelihood of sustained throughput at the lowest latency I can achieve by not saturating the outbound limits on my cable modem comcast has set after 30 seconds (based on my experience of my connection).

I'm sure Hoss or AJ if they disagree will say something (or nothing if they agree probably) but as a rule of thumb this seems to work for me.

--

Laptop:~ douglasbaggett$ iperf -c jedi-rh.cise-nsf.gov -i 1 -t 120

------------------------------------------------------------

Client connecting to jedi-rh.cise-nsf.gov, TCP port 5001

TCP window size: 228 KByte (default)

------------------------------------------------------------

[ 3] local 192.168.1.5 port 58741 connected with 192.12.209.76 port 5001