Hub 2 - upgrade over Archer C3200?

alexfthenakis · ‎02-07-2024

I recently moved to FTTP and got a Hub Two so have used it to replace my TP Link Archer C3200:

https://www.tp-link.com/uk/home-networking/wifi-router/archer-c3200/

The TP Link is from ~2016 so I did this on the assumption that Hub Two is probably an improvement at this point. I think I bought the third-party router early in my Plusnet DSL days because the ISP router was shockingly terrible at dealing with local network traffic (wasn't even a Hub One - something more basic than that).

Never had an issue with the TP Link and it was still going strong, so now I've started wondering whether or not the Hub Two is actually an upgrade. How do the two units compare? Is it a pros/cons list kind of thing or is one of them a clearly superior machine?

I'm capable of interpreting all the data on these things but last time I did it was when I bought the TP-Link, so it would take me a whole day to re-learn what I'm even looking at. Hopefully someone who is more up to speed can make a quick assessment at a glance?

Thanks!

Dan_the_Van · ‎25-06-2007

@alexfthenakis welcome to the community

You can find the specification of the Hub two here https://www.plus.net/help/broadband/router-information/

If it were me I would use the C3200 as it is more configurable, you just need to make sure you set the Internet WAN connection is for PPPoE. WAN port to Openreach ONT

Username: youracountname@plusdsl.net

Password: password used to login to the members centre.

The Hub Two can be stored as a spare as it not required anymore

HTH

Anunnaki · ‎15-06-2024

I would choose the C3200 over the Hub-2 because -

it has QoS which can be tuned to improve your internet latency (gaming ping times)
you can choose different SSID names for each of the WiFi bands

MisterW · ‎30-07-2007

My only concern with the C3200 would be throughput limitations if used on the higher speed FF products.

The C3200 is based on the Broadcom BCM4709A chipset, a dual core 1ghz processor. See some information here

https://openwrt.org/docs/techref/hardware/soc/soc.broadcom.bcm47xx

New Broadcom devices with gigabit Ethernet are supported by the bgmac kernel driver. Unfortunately CPUs on most of these SoCs are too slow to provide 1000 Mb/s routing or NAT. It results in NAT being limited to something around 130Mb/s on BCM4706 and even less on slower units (like ~50Mb/s on BCM4718A1).

To solve this problem, Broadcom developed the proprietary ctf.ko module that watches in-system routing rules and implements NAT on its own. It results in much better performance (even up to 850Mb/s on BCM4706) while breaking things like QoS and advanced firewall rules.

Unfortunately ctf.ko is closed source and there is no open source alternative. For more details see:

Its unknown whether the TPlink firmware includes the Broadcom closed source module (it probably does) but using it may preclude the use of QoS anyway.

@alexfthenakis bottom line, if you have a fast FF product ( 300 or up) and wish to use the C3200 then I'd do a wired speed test just to check its not limiting throughput.

Superusers are not staff, but they do have a direct line of communication into the business in order to raise issues, concerns and feedback from the community.

Anunnaki · ‎15-06-2024

From this review and test - TP-LINK Archer C3200 AC3200 Wireless Tri-Band Gigabit Router Reviewed

here are their WAN <--> LAN throughput test results -

EDIT: I've attached the above image below, if it helps your image viewer to zoom to a readable scaling !.

MisterW · ‎30-07-2007

@Anunnaki thanks for that. That would tend to confirm that the TPlink firmware has the proprietary module.

Note that the test process explicitly states QoS disabled.

Superusers are not staff, but they do have a direct line of communication into the business in order to raise issues, concerns and feedback from the community.

alexfthenakis · ‎02-07-2024

Thanks all,

So in terms of my own real-world speed testing should I just run a basic browser-based broadband speed test with each router and compare? Or do I need to push the limits a bit more than that?

Hub Two has actually been in use for a few months now. The reason I even bothered considering/asking about this is that I just outgrew a WD MyCloud and upgraded to a more full-featured Synology NAS and I'm thinking about the local network file transfer speeds. So far I've only pulled one big chunk of data over and was only getting 40-60 MB/s. That's direct from Synology file station so there's no computer or wireless connection involved - literally just Synology pulling from WD over gigabit ethernet, both devices connected directly to the Hub Two router.

So in addition to broadband speed testing I'll also do a bit of local traffic speed testing: I'll swap out routers to try the next chunk of data migration and see if there's any change. It may just be the hardware limits of one of the NAS units (likely the older WD one) but I'll be curious to know for certain. I'm pretty sure my original purchase of the Archer was because the pre-hub Plusnet router couldn't handle serving up a video file from the MyCloud to Apple TV at the same time as I was browsing the internet on a laptop. That's not a challenge for Hub Two but I am curious to see how the Archer will compare.

Pretty sure Archer still has all my settings in it and all I did was add a '1' to the end of the wireless network name, so I should be albe to swap out fairly quickly. Let me know if there are any particular tests I should try running.

Cheers!

MisterW · ‎30-07-2007

So in terms of my own real-world speed testing should I just run a basic browser-based broadband speed test with each router and compare? Or do I need to push the limits a bit more than that?

@alexfthenakis from a personal POV I've found the CLI version of the Ookla speedtest (https://www.speedtest.net/apps/cli) to give the most consistent results when testing the higher bandwidth products. Obviously tests need to be done on a wired connection to the router when nothing else is using the connection.

The Hub 2 has no QoS options but if you intend to use that functionality on the C3200 , then you need to repeat any tests with QoS enabled to check any effects.

Superusers are not staff, but they do have a direct line of communication into the business in order to raise issues, concerns and feedback from the community.

Anunnaki · ‎15-06-2024

@alexfthenakis if I was doing the internet speed comparisons, I'd choose the waveform test, because in addition to the speed results, it also measures the latency delays and bufferbloat grade of your internet connection.

I'd first test the Hub-2, and save the result (using the "Share Your Results" link under the results), this will give you a baseline to beat both in terms of speeds and bufferbloat grade.

Then swap to the Archer C3200, and test that, saving the result. Hopefully the speeds will be similar to the Hub-2.

Then try QoS by setting the Archer C3200 "Bandwidth Control" rate limiters to say 90% of the previous test result speeds, and see whether you get a better bufferbloat grade without killing the throughput speeds (i.e. dramatically worse than 90% of non-QoS throughput).

Let us know the three test results.

.

alexfthenakis · ‎02-07-2024

OK here's a total overload of test results. I used Waveform purely because it didn't require an install. I didn't actively switch anything on the network off to isolate the connection, but I made sure nothing else was updating/browsing/streaming etc. at the time I ran the tests - so background services may have interfered slightly but things were mostly quiet.

To my untrained eye it looks kind of like a wash between the two routers in terms of total speed. I see how the QoS limits can improve stability but I don't have a real-world sense of how that might affect my experience since I've never really had issues with video calls or anything and I don't do low-latency gaming (as highlighted in the bufferbloat score). I do see how the Archer gives me more control there but I'm not sure what I stand to gain/lose by enforcing those limits.

One thing which probably does sell me on keeping the Archer in service is that once I completed my speed testing I set Synology to pull the next big hunk of data off the MyCloud and it's currently floating around 75-85Mbps which is a decent improvement over the previous copy. The NAS is mostly for backup and overflow storage rather than anything impacting workflow though, so the current migration is one of the few times where those speeds will even really make a difference for me.

On to the results:

I ran the test 3x each pretty much back to back for all of the following configurations:

2015 12" Macbook on WiFi, Same Macbook on a Belkin USB-C ethernet adapter, then my 2020 27" iMac on WiFi (too far from router for wired connection although I'm working on that at present).

I did all three on the Hub Two, then on the Archer, then on the Archer with Bandwidth limits set at 200Mbps down and 40Mbps up.

Finally, I did a run on the iMac only with Bandwidth limits adjusted to 300Mbps and 45Mbps to get a better sense of where the crossover point might be in trading speed for stability.

Results below:

HUBTWO

MB12 WiFi

MB12 Ethernet USB-C

iMac27 WiFi

--

ARCHER

MB12 WiFi

MB12 Ethernet

iMac27 WiFi

--

QoS ARCHER – 40,000kbps up / 250,000kbps down

MB12 WiFi

MB12 Ethernet

iMac27 WiFi

--

QoS ARCHER – 45,000/300,000

iMac27 WiFi

What should I be making of those results? And any tips on what settings I should implement going forward?

Thanks!

Anunnaki · ‎15-06-2024

@alexfthenakis wrote:

What should I be making of those results? And any tips on what settings I should implement going forward?

@alexfthenakis - you've done a great job of systematically providing test results !

My observation is that your results for "QoS ARCHER – 45,000/300,000" is pretty close.

I think you can get improved speeds by running slightly more relaxed active latency scores.

My suggestion would be to -

Incrementally increase your UPLOAD bandwidth limiter rate from 45000, until the test result "Upload Active" consistently shows +4ms or +5ms

Then incrementally increase your DOWNLOAD bandwidth limiter rate from 300000, until the test result "Download Active" consistently shows +7ms to +10ms

This will likely give you a bufferbloat grade "A" - perfect for daily use !

("A+" is a bit overkill and is a waste of bandwidth most of the time)

.

alexfthenakis · ‎02-07-2024

Thanks - I assume I'm optimising based on the machine/connection type where stability matters most? In this case the iMac on WiFi?

59391kbps up and 482500kbps down seems to be the sweet spot for the desktop, but it's leaving the laptop a bit less stable in terms of downloads. I'm guessing this is where I'd use device-specific limits if I cared to do so but that it's probably not worth the effort in my case?

(I never intended to go to single-digit granularity but, since I'd done the upload limit first in isolation I kept tinkering with it once I started on download limit out of curiosity. Weirdly, the jump from 59391 to 59392 corresponds to a change from +0-4ms to +25ms or higher. I wasn't interested in spending the time to get to the same level of detail with my download limit so I think I only got as far as steps of 500kbps.)

Some test results based on current settings here:

QoS 482,500/59,391

iMac 27 WiFi

MB12 WiFi

MB12 Ethernet

Anunnaki · ‎15-06-2024

@alexfthenakis

I would say that's close enough, and much better than what you'd achieve with the Hub2.

Try living with it, and see how it goes.

If it needs re-tweaking in future you now know what to do.

Fortunately you're on "Full Fibre" so hopefully your speeds won't vary (requiring re-tweaking) as happens on FTTC or ADSL.

.

Hub 2 - upgrade over Archer C3200?

Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?

Re: Hub 2 - upgrade over Archer C3200?