Extra Tweaking

[wrongsyntax]

Hi, Im new to these forums, I was wondering if using TCP/IP Optimizer on manual can yield better results rather than doing "optimized", and if so, what settings to change.

My advertised connection speed is 2MB on a Virgin Media cable modem:

« SpeedGuide.net Speed Test Results »
1940 kbps down (~1.94 Mbps, 237 KB/s)↓
185 kbps up (~0.19 Mbps, 23 KB/s)↑
5120 KB downloaded in 21.625 seconds
500 KB uploaded in 22.189 seconds
Tested on: 2008.08.25 13:13 EDT
Tested from: eegna.de
Test Link: http://www.speedguide.net/speedtest/results.php?test=3LMVCZKJUYTJ
Provider: nott.cable.ntl.com
Location: Port Talbot, UK




« SpeedGuide.net TCP Analyzer Results »
Tested on: 08.25.2008 13:09
IP address: 82.1.xx.xx


TCP options string: 020405b40103030101010402
MSS: 1460
MTU: 1500
TCP Window: 128480 (multiple of MSS)
RWIN Scaling: 1
Unscaled RWIN : 64240
Reccomended RWINs: 64240, 128480, 256960, 513920
BDP limit (200ms): 5139kbps (642KBytes/s)
BDP limit (500ms): 2056kbps (257KBytes/s)
MTU Discovery: ON
TTL: 50
Timestamps: OFF
SACKs: ON
IP ToS: 00000000 (0)


Thanks in advance!

[trogers]

Better to set TCP Window to 64240 and uncheck Window Scaling.

[wrongsyntax]

I see, how do you know whether to use Window Scaling for a connection? Do you look at what download speeds you are getting, and then adjust for the value that will give you a BDP limit just over what you are getting?

Thanks for the tip :)

Any others that spring to mind?

[trogers]

I see, how do you know whether to use Window Scaling for a connection? Do you look at what download speeds you are getting, and then adjust for the value that will give you a BDP limit just over what you are getting?

Thanks for the tip :)

Any others that spring to mind?

Window scaling is checked only when TCP Window is to be set higher then 65535.

Using the BDP formula:

TCP Window (RWIN) = 2000 kbps x 240ms divided by 8 = 60000

Latency of 240ms is good enough for surfing from the EU to the US.

Thus, follow recommendation of TCP Analyzer on 64240 to get max. unscaled RWIN which is an even multiple of MSS.

[wrongsyntax]

Oh, I see, I forgot there was a formula for RWIN. Thanks for your help!