Road Runner Cable Modem Tweaking Help (i.e. Lobo!, hehe)

[gadwest]

Hello,

Well I have had my connection tweaked out using Speedguide.net's Windows 98 SE connection and IE browser patch. But I am wondering if I can go faster with better RWIN settings, etc.

I am working with Road Runner cable (Upstate, NY area) on two machines- 1.0GHz 256mb RAM, 900MHz 512mb RAM.

Right now here is my current setup: I am not sure if that RWIN is too high cause it seems to be not as responsive, getting about the same download rates, though.

Could you recommend any better settings for RWIN, TTL, etc. I have downloaded the newest version of CableNut 4.0 and have been modifing values using that.

If you need more info please let me know! What other speed testing sites are fairly reliable that you would like me to try out...

Thanks,

Adam West

TCP options string = 020405b40103030301010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 513920
RWIN Scaling (RFC1323) = 3 bits
Unscaled Receive Window = 64240
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
256960 (MSS x 44 * scale factor of 4)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 20556.8 kbps (2569.6 KBytes/s) @ 200ms
Your RcvWindow limits you to: 8222.72 kbps (1027.84 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 47 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000

[blebs]

Take this Speed Test and highlight the results box, control C to copy and then paste the results in this thread please. Going any higher with RWIN is not going to do anything for you. I believe NY RR has a cap implemented and if so, your not going to get over it no matter what you do. Let's see what the result is from the Speed Test first!

[NotFastEnough]

Where are you in Upstate NY? I used to live between Syracuse and Cortland... move back home to Louisiana about 6 months ago.

Try this speed test...

http://nyc.speakeasy.net/

You might play with different RWINS... see what is best for you.

You could also try the system.ini tweak to dedicate some memory to the NICs IRQ. Works for some, for some it don't...

Good luck, Im sure Mr. Lobo will be along soon.

[NotFastEnough]

Hey Blebs, didn't realize we game the same speed test. Sorry about that

[blebs]

Originally posted by NotFastEnough
Hey Blebs, didn't realize we game the same speed test. Sorry about that

I find that speed test to be pretty darn accurate on download speeds, but it's off a touch on upload!
If I were Lobo, I'd be trying to have a BBQ out and away for awhile. Are there any races on today? Maybe he's watching the races.

[gadwest]

Well here is my results from the test:

Your download speed : 1927230 bps, or 1927 kbps.
Browsers would show : about a 235.2 k/sec transfer rate.
Your upload speed : 160320 bps, or 160 kbps.
Your connection rocks .. above the 1mbit barrier!

I have run this test I believe on this site before at night and got around 2000kbps download and 400kbps upload.

Thanks again,

Adam West

[NotFastEnough]

Lobo will be around, you aren't looking bad now but he might have some other ideas for you.

Good luck!

[shaddaddy]

** Speed 1837(down)/369(up) kbps **
(At least 36 times faster than a 56k modem)
Finish.

thats what i got man, seems you have to find the magic settings for rr. i have talked to people that say they get 2meg/sec downloads at night, so i know they are out there. (btw not mbit)

[NotFastEnough]

Im on RoadRunner, 1Ghz, 256MB of RAM

I get 2Mb down and 356 up or so using just the standard Cablenut settings in the Adjuster... except RWIN at 554800 and add 32 in maxconnectionsper1_0server and 16 in maxconnectionsperserver

I'm pretty slow compared to alot of these guys, but I seem to do OK amongst the RoadRunner crowd

[One Liberty]

Hi guys,

Hey NFE, I checked with the boston Mass, connection that is closer than NYC and these are what I got.

Using a system of 6.4 (partitioned) {4.0 Gig & 2.0 Gig), and 1.6 Gig...192 MB SDRam...PCX1100 Modem

Your download speed : 1429021 bps, or 1429 kbps.
Browsers would show : about a 174.4 k/sec transfer rate.
Your upload speed : 284495 bps, or 284 kbps.
Your connection rocks .. above the 1mbit barrier!

Again Im using the Cable Nut Adjuster Program, but it seems that even in the dead of night this is my average, so I called RR and found out that RR in New England has a "Cap" installed to keep the network systems chugging along.

So for now Im just happy with this, Im sure there could be more, but for now it will do.

thanks Not Fast Enough, you've been a great help!

One Liberty

[NotFastEnough]

I guess their standard caps must be 2Mb down and 356k up, that is what they tell me we have here as well.

[Lobo]

gadwest, On 1 GiG CPU puter, try RWIN of 554800, on 900 try 522680 or 524176 on Cablenuts adjuster, GlobalmaxRcpWindowSize-BLANK, save to registry at bottom and reboot each time you change a value in adjuster

[NotFastEnough]

Hey Lobo, what happend to my cool picture at the bottom of my sig? Is that not allowed anymore?

[One Liberty]

Using a buds computer program and testing a CA location, I got back this read from him

Your Tweakable Settings
Receive Window (RWIN): 65535
Window Scaling: off
Path MTU Discovery: ON
RFC1323 Window Scaling: OFF
RFC1323 Time Stamping: OFF
Selective Acks: ON
MSS requested: 1460
TTL: unknown
TTL remaining: 46

Example 146000 byte download
Actual data bytes sent: 204400
Actual data packets: 140
Max size packet sent: 1500
Max size packet recd: 1500
Retransmitted data packets: 40
sacks you sent: 21
pushed data pkts: 7
data transmit time: 23.646 secs
our max idletime: 3719.6 ms
transfer rate: 6110 bytes/sec
transfer rate: 48 kbits/sec
This is not a speed test!
transfer efficiency: 71%

And I just discovered my Cable Modem is on a splitter with my TV!


That would amount to data loss Im sure.

Any suggestions I greatly helpful, other wise, the numbers are far better than I was given in the beginning.

Thanks Lobo!

One Liberty

[NotFastEnough]

That looks like a test from dslreports.com. If so please ignore the results! Their "teak tests" are full of bs!

I have read that cable modems should be on a "drop" by themselves, but have no experience with that end of it. They installed mine on it's own drop.

Good Luck, Lobo should be around soon.

[One Liberty]

TCP options string = 020405b40103030201010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 256960
RWIN Scaling (RFC1323) = 2 bits
Unscaled Receive Window = 64240
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
513920 (MSS x 44 * scale factor of 8)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 10278.4 kbps (1284.8 KBytes/s) @ 200ms
Your RcvWindow limits you to: 4111.36 kbps (513.92 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 49 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000

Okay the test was taken at 3:15 EST with lobos link, deleted all temp files and did all that he said, then made the cable modem to a direct line with out a splitter on it, then increased the RWIN each time in Cablenut 4.0

This is what I got for my second test.

Going for the third test to see if it will be better.

[One Liberty]

Third Test & Current Time 3:33 EST

TCP options string = 020405b40103030301010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 513920
RWIN Scaling (RFC1323) = 3 bits
Unscaled Receive Window = 64240
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
256960 (MSS x 44 * scale factor of 4)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 20556.8 kbps (2569.6 KBytes/s) @ 200ms
Your RcvWindow limits you to: 8222.72 kbps (1027.84 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 49 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000

Also in "MaxConnections" in Cablenut, I increased it by this fomula which seems to be working

64240 -- bits 0 -- 200 MaxCon. (x0 scalefactor)
128480 -- bits 1 -- 300 MaxCon. (x2)
256960 -- bits 2 -- 400 MaxCon. (x4)
513920 -- bits 3 -- 600 MaxCon. (x8)

Going for broke here and using a scale factor of 10 to see what I get

[One Liberty]

time 3:45 am EST

TCP options string = 020405b40103030401010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 642400
RWIN Scaling (RFC1323) = 4 bits
Unscaled Receive Window = 40150
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
513920 (MSS x 44 * scale factor of 8)
256960 (MSS x 44 * scale factor of 4)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 25696 kbps (3212 KBytes/s) @ 200ms
Your RcvWindow limits you to: 10278.4 kbps (1284.8 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 49 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000

Okay in this test the scale factor was 1460x44*10 and the only notice was the unscaled recieve window which is now 40150...any comments, let me know...going to bed...G'night

One Liberty

[Lobo]

One Liberty, Set Tcp1323 Opts to 1, to allow scaling

[One Liberty]

http://www.geocities.com/one_liberty/screen.html

Heres what is running and the Cablenut 4.0 program as it is on my desktop.

The test I get from your link is this:

TCP options string = 020405b40103030401010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 642400
RWIN Scaling (RFC1323) = 4 bits
Unscaled Receive Window = 40150
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
513920 (MSS x 44 * scale factor of 8)
256960 (MSS x 44 * scale factor of 4)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 25696 kbps (3212 KBytes/s) @ 200ms
Your RcvWindow limits you to: 10278.4 kbps (1284.8 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 49 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000

Any suggestions Lobo, should I try a setting of 1460x44*12??

One Liberty

[Lobo]

I would set RWIN to 53071:D , and no it is not a multiple, if you want that for 900 go to 52268:D

[One Liberty]

So Lobo....

By setting it to the numbers you gave me, what does it perform, currently my pages are lighting fast.

as for the Unscaling Recieved Windows 40180, is having this number lower a bad thing??

I will try those numbers...And here is the results:

TCP options string = 020405b40103030301010402

MTU = 1500
MTU is fully optimized for broadband.

MSS = 1460
Maximum useful data in each packet = 1460, which is equal to MSS.

Default Receive Window (RWIN) = 522680
RWIN Scaling (RFC1323) = 3 bits
Unscaled Receive Window = 65335
RWIN is a multiple of MSS
Other values for RWIN that might work well with your current MTU/MSS:
513920 (MSS x 44 * scale factor of 8)
256960 (MSS x 44 * scale factor of 4)
128480 (MSS x 44 * scale factor of 2)
64240 (MSS x 44)

bandwidth * delay product:
Your RcvWindow limits you to: 20907.2 kbps (2613.4 KBytes/s) @ 200ms
Your RcvWindow limits you to: 8362.88 kbps (1045.36 KBytes/s) @ 500ms

MTU Discovery (RFC1191) = ON

Time to live left = 49 hops
TTL value is ok.

Timestamps (RFC1323) = OFF

Selective Acknowledgements (RFC2018) = ON

IP type of service field (RFC1349)= 00000000



Thanks again

One Liberty

[Lobo]

Test looks fine, that 642400 was to big