Looking at the list here and reviewing your results, here’s a quick synopsis:
1. Your interconnect with the 87U is running at 1 Gb/s, so that won’t be a problem.
2. The Nat Acceleration is good where it is. After some searching I came across the following at
http://pcdiy.asus.com/2014/08/august-2014-the-best-802-11ac-router-rt-ac87u-rt-ac87r/:
CTF(Cut Through Forwarding) is software optimization technique to accelerate NAT.
FA(Flow Accelerator) is a hardware NAT acceleration mechanism design for accelerating wired DHCP and Static IP connections.
Level 1 CTF = CTF
Level 2 CTF = FA + CTF
Conditional implementation details for NAT acceleration are noted below.
Enabling of Adaptive QoS will disable FA.
Enabling of traditional QoS will disable both FA+CTF.
So, as I indicated earlier, the Broadcom CTF, and as it turns out the FA as well is not compatible with QOS. So, it remains to be seen if using QOS and processing all of the data to select a particular data type is faster than just allowing the router to run fast for every data type. That will take some experimentation by you.
3. Your ping time looks like the standard times off of the CGN3. Can you run a ping test to your 87U, with the CGN3 in Bridge mode and just post in the bottom of the results section that shows the ping statistics?
4. The speedtest results aren’t too bad for the 60 Mb/s plan. Is that the peak speed that you see, or is that the final result which tends to back off of the peak speed?
5. Your signal levels and signal to noise ratios are really good for both the downstream and upstream, so those will not be any problem at all.
6. The trace is interesting. I see that you’re using one of the WOW America’s servers as the address. I ran the same address last night and ended up with the same results. I ran it again this morning and ended up with different results. Using the 32 Bit version of WinMTR I ended up with the following results:
|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| INTEL_CE_LINUX - 0 | 101 | 101 | 0 | 2 | 46 | 1 |
| INTEL_CE_LINUX - 0 | 101 | 101 | 8 | 13 | 57 | 11 |
| INTEL_CE_LINUX - 0 | 101 | 101 | 9 | 15 | 41 | 16 |
|so-5-0-3.gw02.bloor.phub.net.cable.rogers.c 0 | 101 | 101 | 16 | 25 | 56 | 28 |
| INTEL_CE_LINUX - 0 | 101 | 101 | 25 | 31 | 57 | 29 |
| ix-12-0.tcore2.CT8-Chicago.as6453.net - 0 | 101 | 101 | 26 | 33 | 112 | 33 |
| if-22-2.tcore1.CT8-Chicago.as6453.net - 0 | 101 | 101 | 52 | 59 | 114 | 58 |
| if-12-6.tcore2.NYY-New-York.as6453.net - 84 | 24 | 4 | 55 | 61 | 69 | 60 |
| if-4-4.thar2.NJY-Newark.as6453.net - 0 | 101 | 101 | 50 | 58 | 109 | 56 |
| if-11-2.tcore2.AEQ-Ashburn.as6453.net - 0 | 101 | 101 | 53 | 61 | 110 | 56 |
| No response from host - 100 | 21 | 0 | 0 | 0 | 0 | 0 |
|________________________________________________|______|______|______|______|______|______|
Deselecting the “Resolve name” selection in the Options and running the test again, here are the IP addresses.
|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 10.0.0.1 - 0 | 100 | 100 | 0 | 2 | 83 | 1 |
| 7.11.164.XXX - 0 | 100 | 100 | 8 | 14 | 92 | 15 |
| 67.231.220.29 - 0 | 100 | 100 | 9 | 15 | 32 | 14 |
| 66.185.81.73 - 0 | 100 | 100 | 17 | 25 | 109 | 22 |
| 24.156.144.182 - 0 | 100 | 100 | 25 | 33 | 132 | 27 |
| 64.86.79.89 - 0 | 100 | 100 | 25 | 32 | 109 | 29 |
| 64.86.79.2 - 0 | 100 | 100 | 53 | 59 | 151 | 54 |
| 216.6.99.45 - 3 | 92 | 90 | 0 | 64 | 74 | 61 |
| 66.198.111.17 - 0 | 100 | 100 | 51 | 58 | 151 | 57 |
| 216.6.87.137 - 0 | 100 | 100 | 52 | 64 | 151 | 58 |
| No response from host - 100 | 20 | 0 | 0 | 0 | 0 | 0 |
|
|________________________________________________|______|______|______|______|______|______|
I’ve cut out the last “No response from host” listings to shorten the list. The final address that you indicated is an AT&T address. The point on the WinMTR results, where there is a loss indicated is a Tata communications address in New York, while the last address on the list is also from Tata, this time in Ashburn Virginia I believe. From that point, there are no results in the trace, which makes it impossible to determine where any other problems might be. Using other online trace facilities ends up in the same situation, where the trace doesn’t complete.
I think, from looking at your results, without seeing the ping times to the 87U, that the problem is a path problem from your location to the WOW server. The difference from Bell to Rogers would be the peering agreements, which would dictate the path that is traversed. Can anything be done about it, that’s a tough one. Maybe Gdkitty can offer some advice on this one.
Lastly, it would appear that you’re not having any issues with your Realtek Ethernet adapter, which I assume is the motherboard adapter. Some people have nothing but problems with Realtek adapaters, and others have no problem what so ever. One thing I would do is navigate to the IPV6 settings on the 87U and disable the Connection type and disable the Enable Router Advertisement, and for your motherboard devices as seen in the Control Panel….Device Manager, navigate to the Advanced settings for the Ethernet and wifi adapters and disable IPV6 if there is a separate enable / disable selection for IPV6.
- With your PC connected to the CGN3 (assuming that its in Gateway mode), log in and go to BASIC…..GATEWAY FUNCTION. Disable the Residential Gateway Function. Save the changes and it should then reboot into Bridge mode, where the bottom two ports, 1 & 2 should then be active. At this point everything should be disconnected from the CGN3.
- With the router unplugged from the CGN3, power it up and run a factory reset using the reset button at the back of the router. Once done power off the router.
- Connect the router to the CGN3 and your PC to the router.
- Power up the router and when its up, log in. The PC should automatically pick up a LAN IP address. If not, you might have to reboot the PC. The connected port LED at the back of the CGN3 should be amber indicating 1 Gb/s connection rate with the router.
- Check the following settings on the router:
- Administration – check to see if the router is in Wireless router mode which is its default mode
- Firewall – Enable - Firewall
Disable - Respond Ping Request from WAN ( ie: set to no)
3. IPV6 - Disable Connection Type
Disable - “Enable Router Advertisement
4. Wireless – WPS – Disable – “Enable WPS”
General – Configure the wireless networks as required
5. LAN – Switch Control - Level 2 CTF selected for NAT acceleration
Jumbo Frames could be enabled if you know that the Realtek adapter can
support it and you can enable that mode in the adapter advanced settings.
6. Adaptive QOS – Bandwidth Monitor – Off
QOS - Enable Smart QOS - off
Note that as I indicated earlier, the QOS and Adaptive QOS are not compatible with the Nat Acceleration. However, at 80 Mb/s as a max rate, it might not make any difference as to whether or not the Nat Acceleration is actually up and running. You might be better off running the adaptive QOS, selecting it for Adaptive Type and using the menu to push gaming to the highest priority. Given that I believe the issue is with the path to the server, I don’t think it will make any difference, but you will have to experiment with it to see if there is any difference in terms of latency or drop-outs from the game.
7. WAN – Enable UPnP – my preference is to set this to “No” which would prevent programs from changing router settings without your knowledge, which can be a security hazard. Some people prefer to leave this up and running. I’ll leave this at personal choice……
8. WAN – Dual WAN –disable
Nat Passthrough – Disable VPN passthroughs unless you need them
That should do it.
Once everything is up and running, run a ping test to the router. You should see constant 1 ms or less response times. Please post the bottom of the ping test showing the statistics. Run a speedtest as well and post the results. From there it will be time to see what the results are when you are gaming with WOW.
Jumping back and looking at a few things..
I think we are looking at TWO things here.
1) The latency with brided/non bridged. Best all and all to bridge, that will remove the tiny spikes when non bridged.
2) This started about latency in WOW alltogether. Reading the first thread again, that you came from Bell.
From my recolection, Bell and rogers, use DIFFERENT peering partners... different routes from here to blizzards servers...
That in itself alone, could equate to differences in latency overall.
@Armega, that is a very interesting result. I’ve been wanting for someone to do that, ping the external address with the CGN3 in Bridge mode. Personal opinion, that is not a good result. Now that is only a one ping test. Can you run a 100 ping test to the external WAN address instead of a pathping? That will provide a better average. Running 1000 for a test would take a while but would provide a pretty accurate average. The average that I usually see, pinging the internal address is 4 to 5 ms. The average that I see running a 100 ping test for the external IP address is 5 ms. I would expect the internal and external ping times to be the same in Bridge mode as they are in Gateway mode, although I was hoping that the Gateway mode latency would be much better. Can you post that ping test result with the last 10 pings just to see the time variance in some of the pings.
You were indicating that the latency with WOW hasn’t changed. I was thinking that you might see that result, given all of the data that you posted earlier. I would think at this point all that you can do is gather up a few traces and WinMTR results and send them off to Blizzard to see if they can’t get something done about the servers in the U.S. that run data back and forth to their realm servers. I reran a WinMTR session to the server you used for your traces last night and interestingly I ended up with the same Level3 servers and had no response to the trace after that. This should show a switch to the AT&T network at some point. The fact that I ended with the Level3 servers is interesting as I ended up with Tata Communication servers this morning. In both cases the traces did not complete.
Lastly, can you edit your earlier post and remove the last two data fields from your IP address, ie substitute the last two fields with xxx.xxx
