| reply to Engineer_TA
Re: [Wired] WRT54GS Limits Bandwidth said by Engineer_TA:If you still have Tomato installed, you could try to change the WAN port to a LAN port and play with it to see if it makes a difference. I successfully re-assigned the WAN port to port 4, and plugged the cable from the modem into port 4.
I can connect to the Internet, but I ran the speed test, and I didn't detect any change.
I have a suspicion that this might be a duplex mismatch between the router and modem. Else the router is simply busted. |
|
| FYI, I started a thread on the Road Runner (Time Warner Cable) forum about hardware that can handle 20 Mbps or greater:
»[TWC] Hardware for Bandwidth 20 Mbps or Greater (Survey) |
|
| Sorry, I'm out of ideas right now. If I think of anything, I'll be sure to post it.  |
|
 Bill_MIBill In MichiganPremium,MVM
join:2001-01-03
Royal Oak, MI
kudos:1
 ·Comcast
·WOW Internet and..
2 edits | reply to NeedForSpeed5
Just some far fetched ideas The WRT54GS v4 is virtually identical to the WRT54GL. No reason in the world it should go that slow. In the stretching-it department...
Just to make sure you've completely reset to defaults?
Any chance you have a bad cable only showing up with this unit? Swap them all just to look for a surprise.
The IP it pulls from your ISP... often linked directly to the MAC address... isn't in a crappy subnet while your other connections are not?
For fun, clone the MAC address of a speedy connection - to make sure the ISP isn't doing something funny with the MAC in that unit.
Like I said... all in the far-fetched department. 
EDIT: These LAN-to-WAN benchmarks were done with a WRT54GS v3. Except for memory size it should be identical to your v4. Notice 24mbps should be NO problem with stock firmware: »Benchmarking WRT Firmware... Some Surprises! |
|
|
|
| said by Bill_MI:The WRT54GS v4 is virtually identical to the WRT54GL. No reason in the world it should go that slow. In the stretching-it department... Just to make sure you've completely reset to defaults? Any chance you have a bad cable only showing up with this unit? Swap them all just to look for a surprise. The IP it pulls from your ISP... often linked directly to the MAC address... isn't in a crappy subnet while your other connections are not? For fun, clone the MAC address of a speedy connection - to make sure the ISP isn't doing something funny with the MAC in that unit. Like I said... all in the far-fetched department. EDIT: These LAN-to-WAN benchmarks were done with a WRT54GS v3. Except for memory size it should be identical to your v4. Notice 24mbps should be NO problem with stock firmware: » Benchmarking WRT Firmware... Some Surprises! You know Bill, that's a great thought. Possible that the ISP is bandwidth limiting based on IP. Never thought for a moment of cloning the PC's MAC with the router and so very easy to do in Tomato. Go to Tomato setup page (login), go to Advanced, MAC Address and select "Clone PC" beside the WAN port. and reboot (if necessary to reboot) and then see what happens. Great suggestion (fingers crossed). |
|
| Hey, I have good news!! My Linksys router has registered a dramatic improvement. Here is one speed test result for this configuration:
PC --> Linksys Router (Wired) --> Motorola Modem
Here is how I did it. I started a thread on the official Linksys forum. Then someone posted a response saying I should set the MTU to 1300. Earlier on this thread, Engineer_TA did advise me to change my MTU, but I guess I wasn't agressive enough.
I took another look at the MTU, which was set at the default of 1500. I tested a MTU range from 1200 to 1300, and found my optimum MTU at 1243. At that optimum, I noticed a speed improvement from 8 Mbps to 14 Mbps, which broke the 10 Mbps barrier.
I made another change to the WAN port speed from Auto to 100 Mb Full, so I can force the duplex setting to that speed. That actually made it worse, because my speed dropped back below 10 Mbps, so I immediately set it back to Auto.
Then I read your suggestion yesterday:
said by Engineer_TA:said by Bill_MI:The IP it pulls from your ISP... often linked directly to the MAC address... isn't in a crappy subnet while your other connections are not? For fun, clone the MAC address of a speedy connection - to make sure the ISP isn't doing something funny with the MAC in that unit. Like I said... all in the far-fetched department. You know Bill, that's a great thought. Possible that the ISP is bandwidth limiting based on IP. Never thought for a moment of cloning the PC's MAC with the router and so very easy to do in Tomato. Go to Tomato setup page (login), go to Advanced, MAC Address and select "Clone PC" beside the WAN port. and reboot (if necessary to reboot) and then see what happens. Great suggestion (fingers crossed). I was skeptical at first, because I couldn't see why the speed should depend on the device MAC address connected to the modem, and even then, only the modem MAC address should be visible to the ISP, or at least I thought.
Then I performed the following experiment this morning. I tried out different MAC addresses (on the Tomato firmware) for the exact same hardware configuration of:
PC --> Linksys Router (Wired) --> Motorola Modem
I ran the speed test 5 times for each MAC address and averaged the results for each trial. Here are the results (in Mbps):
Random Number MAC
Download (average): 21.33
Upload (average): 1.86
Default Router MAC
Download (average): 17.96
Upload (average): 1.84
PC Clone MAC
Download (average): 16.90
Upload (average): 1.85
Random Number MAC
Download (average): 30.62
Upload (average): 1.87
The last trial "Random Number MAC" was another MAC number randomly generated by Tomato, with the impressive average result of 30.62 Mbps.
I will post later explaning what I think might be going on. But I think my ISP throttled my original MAC number at 8 Mbps. |
|
4 edits | 
I've had my ISP do the same thing when I changed the device connected to the router. Usually, cloning the MAC address of the original item fixes it. I have had to call on occasion to the ISP to get them to reset the system to allow the new MAC (new router for example) to run at full speed. I had forgotten.
As for the MTU, it was suggested that you start around 1,400 and then go + or - 10 until it gets better. I think I posted poorly so you might have just went 1390, 1400, 1410 and not any higher or lower. Sorry if that were the case.
What is your MTU set at as you are now running faster than your original post with PC connected directly to the modem?
Also, did you reset your port #4 back to LAN or leave it on the WAN port? I think you can do a NVRAM reset under Administration, Configuration, Restore Default Configuration section of Tomato. I can also try to figure out how to manually turn port 4 back to LAN, lol.
Finally, Kudos to Bill for suggesting it!!!
Great job!
|
|
Bill_MIBill In MichiganPremium,MVM
join:2001-01-03
Royal Oak, MI
kudos:1 Reviews:
·Comcast
·WOW Internet and..
| reply to NeedForSpeed5
I'd like to hear your theory. I wouldn't put too much weight on intentional settings... it sounds like your cable system has a lot of cobwebs or something.  You've found a few things that don't make sense (MTU is hardly EVER that subtle).
Suggestion. Watch the public IP you get (»ipchicken.com). The same MAC may tend to get the same address or it may be random. The IP range you get may correlate to other things - find the bad and good ranges or verify it makes no difference. |
|
TX CS AggieTX CS Aggie
join:2009-07-22
College Station, TX
kudos:1
1 edit | reply to NeedForSpeed5
Upon seeing that you got your problem resolved... good kob by the way... i thought i would play with my routers mac just to see if there was anything going on with my connection
Each one of these results represents and average of 5 samples from speedtest.com
Cable modem: Motorola SB5120
Router:WRT54GSv4 (tomato 1.25)
Default 1st time
DL 8.298
UL .48
PING 46.4
Clone PC 1st time
DL 8.336
UL .48
PING 46.8
RANDOM
DL 8.34
UL .48
PING 51.6
Default 2nd time
DL 8.152
UL .48
PING 47.2
Clone PC 2nd time
DL 8.264
UL .48
PING 46.2
RANDOM (diff mac)
DL 8.272
UL .47
PING 52.4
Default 3rd time
DL 8.214
UL .474
PING 52.2
Clone PC 3rd time
DL 8.31
UL .48
PING 46.2
After doing the 3rd default test I decided it was best to clone my pc mac for best ping and good speed.
But now i am wondering if i want the best gaming performance for playing on my Xbox 360 would it be best to clone the mac of my 360......hmmmmm.........hmmmmm.... the solution to this thread has raised some interesting questions in my head.
and why would a random MAC give be a not as good ping but best DL? |
|
| reply to NeedForSpeed5
Re: [Wired] WRT54GS Limits Bandwidth Unless changing the mac forces a change of IP which in turn messes up the speed, Which to me just means something is hosed on the ISP. Not balanced something odd.
Does this IP throttle or block say p2p. Have restrictions on running servers. then enforce thru throttling? |
|
2 edits | reply to Engineer_TA
Re: Just some far fetched ideas said by Engineer_TA:What is your MTU set at as you are now running faster than your original post with PC connected directly to the modem? MTU of 1243. My original MTU was the default at 1500.
said by Bill_MI:I'd like to hear your theory. I wouldn't put too much weight on intentional settings... it sounds like your cable system has a lot of cobwebs or something. You've found a few things that don't make sense (MTU is hardly EVER that subtle). I think I can be reasonably confident that the MTU adjustment did enhance the router speed, at least a notch above 10 Mbps (12 to 14 Mbs). This is because I can change the MTU without having to reset the modem, thus keeping the I.P. constant.
If I adjust the MTU either below or above a MTU of 1243, then my speeds falls below 10 Mbps, and I can replicate this consistently.
Before my speed rarely exceeded 10 Mbps, and was stuck at 8 Mbps most of the time.
said by Engineer_TA:Also, did you reset your port #4 back to LAN or leave it on the WAN port? I think you can do a NVRAM reset under Administration, Configuration, Restore Default Configuration section of Tomato. I can also try to figure out how to manually turn port 4 back to LAN, lol. I reinstalled Tomato and wiped out all the previous settings, so all my ports are normal now. I'm also using the WRT54GSv4.bin now. |
|
1 edit | reply to no_one
Re: [Wired] WRT54GS Limits Bandwidth said by no_one :
Unless changing the mac forces a change of IP which in turn messes up the speed, Which to me just means something is hosed on the ISP. Not balanced something odd.
Does this IP throttle or block say p2p. Have restrictions on running servers. then enforce thru throttling?
said by Bill_MI:Suggestion. Watch the public IP you get (» ipchicken.com). The same MAC may tend to get the same address or it may be random. The IP range you get may correlate to other things - find the bad and good ranges or verify it makes no difference. There is someone on the Road Runner forum who asked how to change the I.P. address, because the person was banned on some other forum and wants to get around the ban.
My response would had been to reset the cable modem, and the ISP will simply dynamically re-assign you another I.P. address.
It seems with Road Runner it's not that simple. Apparently, the "IP is tied to the MAC address", so it seems to change the I.P. address, you need to change the MAC address as well.
If the speed is indeed linked to the I.P. address, there are two possibilities:
1. There are fast I.P.'s and slow I.P's., meaning the ISP network as a whole is unbalanced.
2. The ISP throttles subscribers based on the MAC address, I.P. address, or some combination thereof.
I'm not to terribly inclined to experiment with whether the I.P. address is tightly bound to the MAC address or not, because frankly I reached my target speed and I don't want to mess that up. |
|
Bill_MIBill In MichiganPremium,MVM
join:2001-01-03
Royal Oak, MI
kudos:1 Reviews:
·Comcast
·WOW Internet and..
| Maybe you've found a result of some creative filtering going on at the ISP. I say that because the MTU reaction you're finding strikes me as artificial - not a normal MTU behavior.
I know what I'd be doing... plotting max speed vs. MTU to get a good picture of what's happening.
Then there's always the problem of putting all your eggs in one basket. Are you using more than 1 speed test site? It's always possible the test site, or the routing path to it, is the culprit making the IP or even MTU sensitivity. |
|
| said by Bill_MI:Maybe you've found a result of some creative filtering going on at the ISP. I say that because the MTU reaction you're finding strikes me as artificial - not a normal MTU behavior. I know what I'd be doing... plotting max speed vs. MTU to get a good picture of what's happening. To clarify, I do not attribute all of the speed gain to the MTU adjustment. I am saying the MTU adjustment provided a slight improvement in speed.
If I were to guess, most of the speed gain came from changing the MAC address of the router to some random value.
said by Bill_MI:Then there's always the problem of putting all your eggs in one basket. Are you using more than 1 speed test site? It's always possible the test site, or the routing path to it, is the culprit making the IP or even MTU sensitivity. That is a fair point. I may need to try some other speed test sites to be sure. |
|
1 edit | reply to NeedForSpeed5
Thanks everyone again for all your help, and a special thanks and major kudos to Bill_MI for his unique insight into the problem and Engineer_TA for instructing me how to set it up.
Here is my theory of what I believe happened. Earlier I mentioned:
said by NeedForSpeed5:I was skeptical at first, because I couldn't see why the speed should depend on the device MAC address connected to the modem, and even then, only the modem MAC address should be visible to the ISP, or at least I thought. This is my rational for my initial skepticism. The Motorola cable modem has its own MAC address and that never changes. Before I can connect up the modem to the ISP, I had to call the ISP and provide the help desk the modem MAC address. That device is the only modem allowed to access the network, and any other modem would be blocked via the MAC address.
But I never gave them the MAC address for my PC or router, since I can swap any device without calling them.
That is why PC cloning did not make any sense to me, because the ISP was not even aware of the MAC address of my PC (or any other device, except the modem).
However, I did not consider, at the time, that the modem itself would reveal to the ISP the MAC address of any device connecting to it, such as a PC or router.
Through the modem web interface, I discovered that the modem registered the MAC address of the connected device in its memory. This registration also granted that one device exclusive access to the modem, and blocked all other devices.
For example, in connecting directly to the modem, I cannot swap one PC for another PC, since the first PC had exclusive access, while the second PC did not.
For me to swap in another PC, I had to reset the modem, thus wiping out the previous MAC address for that device, and make it register another MAC address. Then the modem will allow me to swap in and connect the other PC.
Because the MAC address is stored in the modem, the modem could also reveal to the ISP that same MAC address of the connected device (be it a PC or router).
Let's assume that the MAC address is tightly bound to an assigned I.P. address, then this has a number of implications:
quote:
1. If the I.P. address is assigned based solely on the MAC address of the modem, then the I.P. would never change, unless the modem was replaced.
2. If the I.P. address is assigned based on the MAC address of the device connected to the modem, then every device, with a unique MAC address, would be assigned its very own I.P, and everytime the device is swapped in, the I.P. address would change.
3. If the Internet speed is based solely on the I.P. address, then a swap of a device would also change the speed.
If we accept implications #1 and #3, then the speed would remain consistent, because the modem would rarely change.
If we accept implications #2 and #3, then the speed would depend on what device you attach to the modem. But since a device is identified through its MAC address, then changing the MAC address would also change the speed as well.
However, the MAC address does not even have to be for a real device, but can also be any arbitrary random number.
Because of this, an arbitrary setting within the router of its MAC address will also force the I.P. address to change on the modem, and as a consequence change the Internet speed as well. Q.E.D. |
|
Bill_MIBill In MichiganPremium,MVM
join:2001-01-03
Royal Oak, MI
kudos:1 Reviews:
·Comcast
·WOW Internet and..
| There's some network conventions that get in the way of some of your thoughts. Things like...
The customer must not be able to obtain more than the approved number of public IP addresses at a time. But standard DHCP must be used, because everyone has it. DHCP is designed for giving out many addresses to many network card (MACs) and a popular DHCP operation is to assign the same IP to the same MAC (often called a "sticky" IP-to-MAC table).
So the artificial "MAC bonding" is used. Once 1 MAC has an IP, no other MAC may obtain an IP. To give out an IP to a different MAC takes 1 of 2 things: 1) Releasing (a DHCP function) the 1st MAC or 2) simply resetting the modem so it forgets that 1st MAC. My system used to do 1) but I haven't checked lately. So 2) is the easiest.
Think of that MAC bonding as a modem function but the DHCP server is separate and out in the ISP network somewhere. These things have to work together to insure proper customer service.
At that point, no MAC bonded, DHCP starts from scratch and assigns the IP from a pool and records this bonding ("sticky" IP-to-MAC table). After a while, that 1st MAC gets removed from this table (returned to the "pool" of addresses). This is at the DHCP server.
Enter the IP address assigned. Why are some faster than others? Many reasons are possible which stretches all the way from customer to the speed test site. But it's likely the ISP equipment may simply be expanded and configured in ways that makes the inequality happen. Some networks on older systems, some new, some getting too many customers (imbalanced).
An IP address can also have a history before you get it... is it being attacked because of what the previous owner was doing with it?
Hope this makes sense. |
|
