Slow Internet Speed Test Shown on Adaptive QoS Page

[Tech9]

It depends on what you use Internet for. I have 500/30 and don't need QoS. No heavy uploaders or gamers around.

 

[Ellenswamy]

It depends on what you use Internet for. I have 500/30 and don't need QoS. No heavy uploaders or gamers around.

I guess my heavy uploads would be coming from my cameras only

 

[Tech9]

I guess not. You don't need to stream full resolution all the time. My 6x cameras fit under 2Mbps upload in monitoring mode. When I need to see something in detail, then I can switch to full resolution, recorded or real-time. Waste of bandwidth is an issue many users don't know how to deal with.

 

[Piggie]

Not the best way, if you need to turn off router features to get somewhat accurate measurements between your router and ISP. You don't "work" on the router, but on the clients. Test what speed the clients get - wired or wireless. This is the important part for you.

True, you work on your clients. And if you're client gets adequate download speeds, you may never notice the speed at the router. As you note, a client doesn't measure the speed at the router. It measures its speed at its interface with the network. If you pay for 1gbps internet service and you're running a 100mb switch, your client can only measure a top speed of 100mbs -- which is fine for almost all applications. If you have multiple switches and 50 clients on such a network and you're getting slow speeds at some of the clients, where do measure from? You can check every client, every switch, or you can carry a computer to the network shed, unlock the network shed, plug into the router and then plug it into the modem - or you can measure the speed at the router from any of the 50 clients or from the comforts of your office. If the speed at the router looks funny, you can remotely re-boot it.

The internal router speedtest is one tool. Just like any tool, it has it's uses -- and it's very useful on a snowy day, on a farm, when I don't want to get my feet wet and put on coat.

 

[Tech9]

As you note, a client doesn't measure the speed at the router.

It does, but you need to select the right client for speed tests - wired and with enough hardware, usually a PC. On proper Gigabit ISP lines this client will show you up to 940Mbps very consistently. No need to fiddle with router settings to get the number from a slower than Raspberry Pi hardware.

or you can measure the speed at the router from any of the 50 clients or from the comforts of your office

I can do that on my x86 firewall, if needed. You may have to play with router settings because of home routers hardware limitations. The idea is okay, but it needs more resources than what's available in home routers. The more things you run, the more inconsistent and inaccurate results you'll get.

 

[smodu]

This is all very interesting to follow, and perhaps the good folk here can help me understand something:

I have the RT-AX88U with the latest stock firmware, and have just been moved onto FTTP 900. My household computers are all Macs with no ethernet ports, so the only device based testing I can do is by wifi - I get anything from 550-800 depending on device and room, which seems fine. On the internal router speedtest, after checking a few servers and discovering that the best one was not the closest, I can now see speeds of around 850-910, which also seems fine. I'd like to use Aiprotection, on the assumption that it does add some security (but I don't use QoS). But when I have AiProtection on, the internal speedtest drops to around 550 every time - I have checked multiple times, and always higher internal test result when aiprotection is off. The devices mainly seem to stay the same, but it's hard to be sure because of the wifi fluctuations...

My question is whether when I turn on aiprotection the speed is really dropping in a way which could affect the devices? Or is it just the result showing on the router that changes because of the aiprotection, but the actual speed it delivers to the house remains as high as before same?

 

[OzarkEdge]

The devices mainly seem to stay the same, but it's hard to be sure because of the wifi fluctuations...

If wireless client speed tests "stay the same", then you could assume the router speed test is affected by AiProtection and forget about it. Otherwise, wire a fast PC to the router and test that.

OE

 

[smodu]

Thanks, but the wireless clients' speed does not seem to be a reliable test as they constantly change. And I can't wire a pc unless I go out and buy one.. we just have two Mac laptops without Ethernet ports.

I'm curious about how the router's internal Speedtest works, and whether the fact that the AiProtection function causes slower internal results has any effect on the real speed delivered from the router.

 

[Tech9]

AiProtection does affect maximum speeds, but higher than 550Mbps. Perhaps about 750Mbps on Gigabit line. This is what the hardware can do. You have to decide what's more important for you - speed or eventual added security. In real life it doesn't matter if you get 500Mbps or 900Mbps to your clients, unless you constantly download large files from Internet. This speed testing obsession goes away over time.

 

[OzarkEdge]

Thanks, but the wireless clients' speed does not seem to be a reliable test as they constantly change. And I can't wire a pc unless I go out and buy one.. we just have two Mac laptops without Ethernet ports.

I'm curious about how the router's internal Speedtest works, and whether the fact that the AiProtection function causes slower internal results has any effect on the real speed delivered from the router.

I understand Macs' lack of ports... that problem is spreading to PCs.

The advice remains to use a wired, fast PC/Mac speed test for comparison... at the fiber ONT, too, if possible.

I would be concerned about 'wireless client speed tests constantly changing'... a wireless Mac ought to yield steady speed test results if the router speed test results are steady.

OE

 

[smodu]

I would be concerned about 'wireless client speed tests constantly changing'... a wireless Mac ought to yield steady speed test results if the router speed test results are steady.

Perhaps it's because of other wifi traffic at same time, or other things working at same time on the laptop, but no two tests ever come out the same.

AiProtection does affect maximum speeds, but higher than 550Mbps. Perhaps about 750Mbps on Gigabit line. This is what the hardware can do. You have to decide what's more important for you - speed or eventual added security. In real life it doesn't matter if you get 500Mbps or 900Mbps to your clients, unless you constantly download large files from Internet. This speed testing obsession goes away over time.

Yes, I'm probably a little too obsessed with the speed having just had the fttp installed...! I expect to care a little less very soon, given that even at the 'lower' router speed all the devices seem to be getting above 500 each. Key thing, I guess, is that without AiProtection I am seeing close to 900, so at least I know that I'm getting what I paid for from the provider. As for the Aiprotection, I'll monitor what it catches over the next few weeks, and then decide whether to stick with it. Views on this forum seem mixed (and people seem to have very strong views!)

 

[Tech9]

and then decide whether to stick with it.

Sounds like a plan. In my testing with different Asus routers I found the routers are faster and more stable when TrendMicro is disabled.

 

[OzarkEdge]

Perhaps it's because of other wifi traffic at same time, or other things working at same time on the laptop, but no two tests ever come out the same.

Yes, I'm probably a little too obsessed with the speed having just had the fttp installed...! I expect to care a little less very soon, given that even at the 'lower' router speed all the devices seem to be getting above 500 each. Key thing, I guess, is that without AiProtection I am seeing close to 900, so at least I know that I'm getting what I paid for from the provider. As for the Aiprotection, I'll monitor what it catches over the next few weeks, and then decide whether to stick with it. Views on this forum seem mixed (and people seem to have very strong views!)

AiP here blocks malicious sites according to TM. You can also manually test URLs on this TM page: Trend Micro Site Safety Center

OE

 

[L&LD]

@smodu, you don't need a wired PC to test with. Use the network equipment you have and use now.

The reason that testing isn't needed is that the internal speed test is known to drop with AiProtection on. The speeds to the client devices shouldn't be significantly affected though (as you've already stated).

Turn off your 'testing mode' and just enjoy the upgraded speeds and the benefits AiProtection offers.

There is nothing wrong with your network as-is. The 'fault' lies in the grossly underpowered router hardware we're dealing with (still), for the features available.

 

[Piggie]

This is all very interesting to follow, and perhaps the good folk here can help me understand something:

I have the RT-AX88U with the latest stock firmware, and have just been moved onto FTTP 900. My household computers are all Macs with no ethernet ports, so the only device based testing I can do is by wifi - I get anything from 550-800 depending on device and room, which seems fine. On the internal router speedtest, after checking a few servers and discovering that the best one was not the closest, I can now see speeds of around 850-910, which also seems fine. I'd like to use Aiprotection, on the assumption that it does add some security (but I don't use QoS). But when I have AiProtection on, the internal speedtest drops to around 550 every time - I have checked multiple times, and always higher internal test result when aiprotection is off. The devices mainly seem to stay the same, but it's hard to be sure because of the wifi fluctuations...

My question is whether when I turn on aiprotection the speed is really dropping in a way which could affect the devices? Or is it just the result showing on the router that changes because of the aiprotection, but the actual speed it delivers to the house remains as high as before same?

With aiprotection in you will see a small drop in bandwidth. However it isn't the 400+ drop shown by the internal speed test. As previously mentioned, the internal speed test isn't accurate when running traffic monitor and ai protect.

 

[Ellenswamy]

Honestly if you want good protection make a Pi-hole and force all devices through it. You can be sure what exactly your blocking also which is a hide plus.