Right, the speed seems to be fluctuating. But the fact is, once the issue started, it was occurring right away. But the video file was 5GB and on SD videos, it was still working (so again, pointing to a bandwidth problem)
As to my infrastructure:
Modem to connect to the internet (garage)
One of the 2 TPLINK Deco W7200 mesh units connected to the Modem (garage)
Moca Adapter connected to that unit (garage)
Moca adapter on the other end, to which the 2nd mesh unit is connected (living room)
Vero 4K+ connected to that 2nd unit (living room)
All living rooms units are in the same location within a wooden TV stand.
The Synology NAS is DS 920+, running on the latest DSM 7. It’s in the basement and also uses a Moca adapter. But I have seen that in the basement, there are 2 coax outlets, and just one coax cable going to the basement, suggesting that there is a splitter somewhere, and that splitter must reduce the gain to a level that makes transfer speeds very low. I will look for that splitter.
Most cables are Ethernet cables Cat 5 - I have one or two Cat6 I think.
I will run tests on the wifi network now (or when my son is done watching TV, which could be a while)
Regarding the iperf3 test: In normal mode the server receives, the client sends.
If the iperf3 server is running on your Syno, a more realistic scenario is to reverse the transmit mode means run iperf3 -R -c ... on the Vero, so the server is the sender and the client the receiver.
A mnemonic: the side showing the cwnd (congestion window) and retry columns in the output is the sender.
You have a mesh AP connected to a moca adapter that is connected to a moca adapter that is then connected to another mesh AP? Are these mesh access points such that they can’t see each other at all or else configured so they will not mesh with each other so they don’t form a network loop between them and the coax connection? Also have you tried rebooting all of the moca adapters? It isn’t unusual for them to need a bit of manual intervention from time to time in order to bring back their speed and stability. Also note that if you have any other moca devices (cable or satellite tv for example) in your house there is a limit to how many can exist at the same time.
@darwindesign I did experiment a bit with splitters, and indeed if the signal is too weak, no ethernet connection can be made… the transfer speed issues are likely due to that.
I’ve restarted the Moca adapters and ran a test for 100s - there were a few peaks at 938Mbps but on average it was
I will remember to restart the MoCa adapters once in a while, and make sure to test with iperf3.
And if it keeps on going, I’ll move the NAS out of the basement and find another location, probably next to the main router in the garage. Not ideal, but… it’s a compromise either way.
Devices can all see each other on the network. The MoCa should be transparent to either Mesh unit. But the NAS and the Vero4K+ are connected to different Mesh Units, yes.
Also, I’d suggest cleaning the MOCA coax connections (sockets/plugs) with Iso-propyl alcohol from time to time. Make sure the coax cable side of things are clean and tight. They can be prone to oxidisation which can interfere with stabiity, regardless of the manfacturer/sales claims (just like old analog TV aerial cabling).
Just to clarify what I was talking about. If the mesh unit in the garage is talking to the mesh unit in the house over wifi and both of these devices are plugged into moca adapters then you have what is called a “network loop”. This can cause all kinds of problems with the speed and reliability of your network traffic.
Oh, I see! Yes, that’s why I disabled the wifi on the vero4K+. I haven’t tested with ‘wifi only’ yet. I might do that if I have issues with transfer speeds.
Be careful when you add the Mesh Wifi, as @darwindesign has aluded too, don’t put the Mesh backhaul across the Moca cabling, use the Mesh’s Wifi backhaul capability instead.
Also note the differences between routers and switches - why is there a router on the Vero4K+ do you mean a switch ?
I’m not sure if I have correctly conveyed to you what exactly I’m talking about. On the Vero it will only connect to your LAN using the wireless or the wired connection by default. What I’m talking about is that the “mesh” in mesh routers/access points typically means that these devices not only talk to client devices wirelessly, but they also wirelessly talk to each other forming a network bridge from one unit to the next. You can think of it as if you ran a cable from one to the next. The loop would be where if these two units of yours are “meshed” together then there is a (wireless) connection from mesh device one to mesh device two and from mesh device two to your moca link and from your moca link back to mesh device one. There is a network path that circles back on itself which is not allowed in this type of networking. So, if that is how it is configured then the solution is to either cut the wireless link (backhaul) between the two mesh devices so they are not longer working as a “mesh” but rather just two independent access points, or else drop the moca link in your garage. Whichever connection type that provides the faster and/or more stable network connection is probably the one you will want to keep.
There is a router before the Vero4K so that devices like my phone can have a good wireless signal in the living room, or upstairs or in the backyard. Maybe it would be good enough if there was just 1 unit, but I’ve had issues in the previous house where I often lost connectivity and didn’t get updates for long stretches of time. I know WiFi routers have made tremendous progress in the past five years, so maybe I don’t need the living room unit. I also thought about just getting a very long ethernet cable and driving it through the house … aesthetics be damned, I’d avoid headaches.
Yes you are right - removing the 2nd mesh unit (so, vero4K connected to the Moca directly) brings speeds up to 930Mbps.
However my phone’s speed goes from 500Mbps to (5 to 40) Mbps in the living room (I’ve tested with the 2nd mesh unit up and without, I get similar results). I’ll probably keep that 2nd unit connected to the MoCa adapter… (I am aware that I’m now testing the bandwidth from the WAN, but that’s high enough and I can see that internal speeds are the limiting factor).
[update] I tested my vero4K+ again, connected to the 2nd mesh unit connected to the Moca (as it’s been since I set it up) and now I’m seeing speeds of 500+Mbps. Maybe I need to restart the MoCa and the router periodically?.. Are these devices moody?
MoCa tends to be regardless of who makes the adapter. This is completely normal. An access point and/or router sometimes can be as well. I also think MoCa adapters are usually (always?) simplex and shared bandwidth so you may find benefit to running ethernet cables to as many locations as is viable is helpful to not just the device that the cable was run to, but it would free up bandwidth for anything still on the MoCa network as well. Along this line you may find that disconnecting the MoCa in the garage so that link goes modem>mesh AP>mesh AP>MoCa may give better all around speeds if you can establish a good wireless signal between the two.
Carefully reading the setup and the Deco W7200 manual, I wonder whether
Is the second TP-Link using the MoCa-backbone at all or does it communicate with the first TP-Link in the garage via Wifi, only? So what happens when you unplug the ethernet cable from the second TP-Link to the Moca adapter?
Your test using a direct connection to the MoCa-adapter seems to prove that there is an influence/life of their own with these TP-Link devices you don’t expect/understand. Reading the manual it is not clear to me what happenes if you connect two of the Decos with an ethernet cable.
I suggest you put a simple gigabit switch between the MoCa adapter and the second Deco device and connect the Vero4lk+ and the TP-Link device to the switch. As a result, the media player should be decoupled from these TP-Link devices when larger amounts of data are transferred.
The coax cable infra structure is a shared media and other device commication will have an impact for the bandwidth between the NAS and the vero: E.g. if your xbox and the first TP-link in the garage use a larger amount of bandwidth for some file transfer, this will have an influence for the available bandwidth between the NAS and the Vero. That could also be a reason for the fluctuating measurement results.
You are right! A gigabit switch should resolve the issue. Why didn’t I think of that…
About your first point, I did try to disconnect the ethernet cable from the 2nd TP Link and let the units connect via WiFi; it wasn’t better.
About the behavior of the 2 units, it shows on my app when they have detected they’re connected via Ethernet. There is no option to configure anything, but I trust that it stops the wireless backhaul at that point.
About the overall bandwidth on coax - thanks; indeed I was aware of that, and I made sure the Xbox was off when testing.
you guys are the best.