Netflix devices cause wired LAN ping spikes

[Hyper_Eye]

I have had this problem for quite some time. I've posted about it in IRC a couple times over the last few months. My configuration is as follows

24-port patch panel -->
24-port Netgear unmanaged gigabit switch (GS324) -->
ASUS RT-AC88U running merlin 380.68_4 -->
SB8200 cable modem with 500/50 service

I have most of my house wired and connected to the patch panel. Each terminated port of the patch panel is connected to the corresponding port on the switch. The switch is connected to the Realtek (second switch) side of the RT-AC88U. I have a second gigabit device connected directly to the RTK side. I have a few half duplex and 100mbit devices connected to the first set of ports on the router.

The devices connected to the switch are a mix of gigabit and 100mbit devices. One of the rooms has a 4-port gigabit switch connected to one of the GS324 ports. Everything else connected to it is a device including computers, game consoles, smart televisions, smart blu-ray players, etc. The switch and router are correctly negotiating a gigabit connection. I have used iperf to test throughput across devices on the switch, between a device on the switch and the router, and devices on different ports of the router. Everything is as expected with TCP transfer averaging at around 950mbit/s. My internet connection tests show that I am getting the rated speed.

Now for my problem. When one of the smart devices (television, blu-ray player) is used to watch Netflix my latency from any device on the switch to the router gets flaky. This is a common occurrence as I have kids that watch Netflix in their bedrooms. It causes issues with the internet connection. Gaming is flaky. Loading websites is hit or miss. My speed tests are terrible. It's a real problem.

Here is an example of ping to the router when one of the kids are watching Netflix:

64 bytes from 192.168.0.1: icmp_seq=226 ttl=64 time=0.208 ms
64 bytes from 192.168.0.1: icmp_seq=227 ttl=64 time=0.204 ms
64 bytes from 192.168.0.1: icmp_seq=228 ttl=64 time=0.177 ms
64 bytes from 192.168.0.1: icmp_seq=229 ttl=64 time=0.172 ms
64 bytes from 192.168.0.1: icmp_seq=230 ttl=64 time=39.3 ms
64 bytes from 192.168.0.1: icmp_seq=231 ttl=64 time=15.3 ms
64 bytes from 192.168.0.1: icmp_seq=232 ttl=64 time=29.2 ms
64 bytes from 192.168.0.1: icmp_seq=233 ttl=64 time=14.7 ms
64 bytes from 192.168.0.1: icmp_seq=234 ttl=64 time=6.56 ms
64 bytes from 192.168.0.1: icmp_seq=235 ttl=64 time=0.201 ms
64 bytes from 192.168.0.1: icmp_seq=236 ttl=64 time=13.5 ms
64 bytes from 192.168.0.1: icmp_seq=237 ttl=64 time=0.174 ms
64 bytes from 192.168.0.1: icmp_seq=238 ttl=64 time=0.202 ms
64 bytes from 192.168.0.1: icmp_seq=239 ttl=64 time=0.177 ms
64 bytes from 192.168.0.1: icmp_seq=240 ttl=64 time=0.191 ms
64 bytes from 192.168.0.1: icmp_seq=241 ttl=64 time=0.188 ms
64 bytes from 192.168.0.1: icmp_seq=242 ttl=64 time=27.9 ms
64 bytes from 192.168.0.1: icmp_seq=243 ttl=64 time=0.186 ms
64 bytes from 192.168.0.1: icmp_seq=244 ttl=64 time=0.200 ms
64 bytes from 192.168.0.1: icmp_seq=245 ttl=64 time=37.4 ms
64 bytes from 192.168.0.1: icmp_seq=246 ttl=64 time=38.1 ms
64 bytes from 192.168.0.1: icmp_seq=247 ttl=64 time=0.192 ms

While this is going on I can ping between two devices on the switch and they do not exhibit this problem. If I ping the device that Netflix is playing on the ping is much lower. If I ping a device on another port of the router from the switch I see this problem. If I log into a device connected directly to the router I can ping the router without this issue and I can ping between two devices that are connected directly to the router fine UNLESS I connect the switch to one of the first four ports on the router. Then this problem affects other ports on the router.

The only thing that seems to resolve it is if I pull the device that is playing Netflix from the switch and plug it back in. Then my ping returns to normal and the device continues playing Netflix. This is temporary though. The next time they turn one of these devices on to watch Netflix it will start again. I know this is not triggered by bandwidth usage. Netflix is barely putting a dent in the usage. I can run a full gigabit stream across the switch and router using iperf and my latency is normal. This is specific to these smart devices. I have noticed that I don't have this problem with my son if he uses his PS3 for Netflix instead of his TV. One difference between those two devices is that the PS3 has gigabit ethernet.

I have my network closet built to accommodate this configuration. I do not want to add another switch to the mix. I shouldn't have to. Does anyone have any ideas as to what could be causing this problem and how I might go about debugging it? I can install entware if it has tools that will help. Does anyone else have a similar problem? So far I have tried numerous versions of merlin's firmware as well as the official ASUS firmware. I have tried disabling NAT Acceleration and STP. The problem is consistent.

Thanks for any suggestions.

 

[elorimer]

Well this won't be all that helpful, but to get it going: when I've looked at netflix streams, they are heavily cached, so they will be downloading a lot to fill the buffer, then stop, then pick it up again, and stop. so it is choppy.

Then, just troubleshooting, I would simplify all of this down to one of these netflix devices, the router, and something else. Then incrementally add a complexity. Few of us I think have a patch panel, which is a great help I think. Leave out all the followon switches first.

 

[Vexira]

Are you by any chance running qos, have you tried disabling spanning tree protocol?

 

[Jack Yaz]

Are you by any chance running qos, have you tried disabling spanning tree protocol?

I have tried disabling NAT Acceleration and STP. The problem is consistent.

OP has tried STP already

 

[Vexira]

sry I missed that thnx for bringing that up.

 

[Csection]

Try disconnecting all those slower devices, then run Netflix. Don't just turn them off, but disconnect them.

 

[Hyper_Eye]

Are you by any chance running qos, have you tried disabling spanning tree protocol?

QoS is disabled. I have also disabled STP without any change.

Try disconnecting all those slower devices, then run Netflix. Don't just turn them off, but disconnect them.

I have been taking steps such as this but it is hard to debug the problem with this approach. As I said before the problem vanishes when the Netflix device is unplugged and plugged back in to the switch. I have not been able to replicate the problem intentionally. I just have to wait for it to happen again which may be the next day. I was hoping for a suggestion as to traffic monitoring I might be able to run on the router to see what is happening on that LAN port to trigger these moments of high latency. I am going to continue working to determine if there are steps to reproduce the issue. I have not ruled out the switch itself. It is fairly new but I have seen in the Netgear forums that their switches sometimes have compatibility issues when used with certain routers or devices. I am considering buying a TP-Link TL-SG1024. For now I will try moving a computer and some of the Netflix devices over to an 8-port Monoprice gigabit switch I have and see if the problem occurs with that.

 

[Vexira]

try enabling adaptive qos with manul bandwith see if that helps, is your modem in bridge mode? Cause it woubds like a bandwith allocation issue.

 

[Hyper_Eye]

try enabling adaptive qos with manul bandwith see if that helps, is your modem in bridge mode? Cause it woubds like a bandwith allocation issue.

The SB8200 does not have routing features. It is a straight customer owned cable modem. The WAN address on my router is my public IP. I don't think the modem is part of the problem. Devices that are hooked directly to the router are unaffected. It is not a bandwidth problem. I am pretty sure it is a problem with the switch, the router, or a combination of the two.

 

[Vexira]

ahh ok it was just a suggestion to see if qos might help, worth a try at least that way you can see if it helps or not, have you checked the cables from the pannel incase any of them have faults same with the cables going to and from the switch, does the switch have a manged interface or is it unmanged?

 

[elorimer]

I am considering buying a TP-Link TL-SG1024. For now I will try moving a computer and some of the Netflix devices over to an 8-port Monoprice gigabit switch I have and see if the problem occurs with that.

A very little bell rang. A few years back I had a problem with an Xbox 360 acting as an extender for MC. The link was Computer (1GB) to router (1GB) to switch1 (1GB) to switch2 (1GB) to Xbox (100mps). The Xbox would disconnect and complain that the network speed wasn't enough, but it would test out ok. The problem was that switch2 (an older Netgear 5 port) had a different buffer size than switch 1, a problem that went away when I replaced it with a TPLink GB 8 port switch.

 

[Csection]

A very little bell rang. A few years back I had a problem with an Xbox 360 acting as an extender for MC. The link was Computer (1GB) to router (1GB) to switch1 (1GB) to switch2 (1GB) to Xbox (100mps). The Xbox would disconnect and complain that the network speed wasn't enough, but it would test out ok. The problem was that switch2 (an older Netgear 5 port) had a different buffer size than switch 1, a problem that went away when I replaced it with a TPLink GB 8 port switch.

This is why I suggested to disconnect devices and only run the Netflix boxes directly from the router. Might be a device slowing down the rest of the network and causing the entire problem(Ethernet problem).

 

[Vexira]

id bypass the panel.and the switch