Mesh system to replace powerline advice

Ged

Need help please. I have 1200 speed TP Link powerlines which I use with a Virgin hub and recently had an upgrade to 200Mb which Speedtest confirms. Unfortunately, whilst only a small, 30 years old house, a 1200 Mb AP connected to the homeplug only connects to the Virgin Router at 76 Mbs. I have tried several of the same speed homeplugs with the same result, whilst testing using a long CAT  cable from the AP to the homeplug connected to the Virgin router gives the expected 200 Mbs.
I have concuded either  a CAT cable or mesh is the solution to the problem, however the cable is not easy to install and would be interested to know if using a Tenda MV6 would ensure a gigabit connection for an 8 port switch in the remote room please?

mickevh

Wi-Fi is inherently unreliable and unpredictable as all Wi-Fi locales differ (both in geography and over time) so no one will be able to say for sure whether any Wi-Fi solution will "work" in any given scenario. Even when armed with some of the best survey equipment, software and expertise sometimes weird things happen. Wi-Fi is "just like that." You basically have to make the best judgement you can and then "just try it" and see what happens. A "proper" cabled ethernet link will be much more reliable and quicker.

A connection to a switch upstream of a Wi-Fi bridge will almost certainly be Gigabit if you use a Gigabit switch and equipment. In ethernet infrastructure, each lobe of the network (a lobe is cable with an active "thing" on each end) negotiates it's "Link Rate" (ever erroneously called "speed") independently of all others.

So (for example) you could build...

R---A~~~B---S

R == Router, A~~~B is a Wi-Fi bridge (or "mesh" link) and S == switch and any Gigabit devices connected to the switch (including the "B" Wi-Fi bridge node if it has GBit NIC) will get a Gigabit link.

But you won't observe Gigabit performance if you run an Internet seed test. Let's say for the sake of debate that the A~~~B Wi-Fi link "only" runs a 40mbps, (and is the slowest link in the pathway between Internet speedtest site and test client) then a speedtest will "only" observe 40mbps. But that doesn't mean the switch "isn't working at Gigabit rates" - it is.

Ged

Thanks for the reply. Just an update, the mesh arrived but is currently unopened as I decided to try another socket for the primary homeplug and the "speedtest" increased.
I can check what speed this tablet connects to the Virgin  router and AP, both connect at 433Mbs (? Link rate?) . However, connecting to the Virgin router gives 200 Mbs internet speed test and the AP now has increased to 120Mbs for internet speedtest. Apart from the internal network rate drop (? link rate via homeplug to AP?) I don't think its worth opening up mesh and will return it for refund.

mickevh

It can be a bit confusing for newbies, but the "speed" of networking technologies means different things depending on context.

When you look at the spec. sheets and cockpits/control panels of the kit, it usually report the "link rate" (literally the rate at which the local link is connected) whereas speedtest etc report the end-to-end average throughput along an entire pathway (strictly speaking speedtest doesn't measure the speed of anything, it measure the time it takes to transfer a measured amount of data and computes a statistical average.)

It's rather like the difference between the speedometer and odometer (or trip computer) in a car. Both are reported in the metric of "miles per hour," but the speedometer and trip computer "mean" different things.

Then there's other horrors (like duplex modes, traffic levels, contention, etc.) to worry about just to muddy the waters further.

Unfortunately, it's all rather complicated and newbies can tie themselves in knots worrying about it.

The 433mbps you cite for the router connection is indeed a "link rate" measure, but your wouldn't expect to see that if you speedtested it, even in ideal conditions.

Speedtest effectively tests the slowest "hop" on the pathway between the device you run the test on and the speedtest server you run the test against (of which there are many to choose.) In olden it always used to be one's Internet link that was the slowest hop and thusly speedtest effectively tested your Internet speed. But with faster and faster Internet links, this is not always the case. For example, the laptop I'm using to type this has a pathetically slow 76mbps Wi-Fi NIC so if I speedtested using your 200mbps Virgin ISP link, the test would be curtailed by my laptop's Wi-Fi performance.

So the take home is to bear in mind that speedtest is an end-to-end throughput test, whereas the kit cites different numbers (the link rates) in the hype and the two are not the same thing even though both report in "bit per second."

Professionals do use things like speedtest, but we know "what's up" with it and only really use it as a rough guide (ie order magnitude indicator) rather start stressing that link says "X" and speedtest only reports (let's say) "70% X." Ultimately, if it's working well enough for ones needs, I tend not to worry about it and get on with my life.

Ged

Indeed, agreed, it became the latest project which upon reflection is not something to worry about.. hence think it will be going back for a refund.