Hi guys! Im new here.
Im looking for a solution to limit the bandwidth by consumption. My wish is that the .1/24 device consumes 35mbps/5mpbs (down/up), the .2/24 device 70mbps/15mbps and so on. I am going to buy an Archer C6 or C80 at the end of the month (as my wallet can) and I am observing that the TP Link firmware no longer has this function like years ago.
Please I need help. Thanks for reading me.
PD: This week an ISP is going to install 300mbps/30mbps for me. I will finally put aside my obsolete 10mbps hahaha
With 300/30 MBit/s and QoS, you should have a good read at:
While I can understand where you're coming from, this
is not quite what you're looking for either, nor easy to implement.
 QCA9563 in rev 1.x, 2.x
 mt7621a in rev 3.x
 Hint, it isn't supported - and never will be. Unsupported SOC and not meeting bare minimum system requirements.
 mt7622bv, filogic 820/ 830 and ipq8072a+ would probably cope with that, but without much of margin for future speed increases; details will have to be tested.
This looks to me like a solution looking for a problem... Wait until you have the new speeds available, and use some QOS to distribute traffic evenly among your devices. Then, if you still have issues with traffic, look for a solution to your actual problems.
I think nft-qos allows per device limits if you really require that. However as @eduperez indicates sqm-scripts with cake can be configured to equitably share your link capacity among all active machines in your internal network.
So .1/24 will get 100% of capacity if nothing else wants to use the network and say .2/24 also has greedy flows, both machines will get 50% of capacity. If one host actually does not use its 'fair' share of capacity that capacity is given to those flows that can use that capacity.
Unlike your desired 'fixed configurable' limits per Internal IP address this 'equitable sharing' model works without the need to manually configure rates for each host. Often that solves most of what people want to use per-device limits for, for those cases it does not, one traditionally needs to built one's own hierarchical traffic-shaper tree with a leaf shaper for each device needing special limits. Nft-qos approximates this, but I have never used it myself, so all I can recommend there to google it....
Oh, I understood your request, all I wanted to say is that if you are not explicitly requiring fixed limits per device, then cake's per-internal-IP-fairness mode might offer something that is already useful, if only as a stop-gap measure while you research/test finer-grained control measures.
As I mentioned, in the past, a number of people (certainly not all) asking similar questions actually found that per-internal-IP-fairness was good enough for their use-cases.
BTW, 1/0.2Mbps is pretty harsh for even "web browsing", as it turns out that up to 10-20 Mbps increasing the access rate actually helps in reducing page loading times (reducing access latency always helps). But that is a different issue.
Good morning Sir.
Let me make a few things clear.
It is true that at the moment my knowledge in networks is of a "beginner" level (I am not ashamed to say it). But one thing is my "academic" level and another thing is that I create "imaginary problems".
I came to this forum of this important firmware to seek guidance if I can perform some functions of commercial brands of router models "N".
I understand that you infinitely surpass my networking knowledge but it would be good if you avoid having the arrogance to know what someone else wants or that you know their life.
I send you greetings and thank you very much for your response. It is not necessary to answer me again. I have already understood you with your attitude.
Please chill a bit. This is a volunteer forum, no one owes you a response, let a lone a response that solves your problem. The best way to deal with responses you deem unhelpful is to politely ignore them.
click at own risk....
Now, @eduperez actually brings up an important question, what do you want to achieve with the bandwidth limit? Because there might be different way to achieve your higher level goal.
Good morning, sir.
Yes, I am relaxed. Maybe he didn't use the right words. I also know that no one is obliged to give me a 100% accurate answer or possible ways of solution.
I understand that I am going to have 300mbps which is awesome. The problem is that there are three of us in this house that are going to pay for the service and we really want to limit the speed (plus there are kids).
So, for example the idea is:
Person1: 100mbps/10mbps (down/up)
Yes, I also know that I will not have 100% of my ISP contract. They are just round numbers.
Anyway, I find it very curious that a function of N routers, old hardware, is so difficult to do on new equipment and "current" hardware.
It often takes two to tango, so maybe consider whether did your best trying to defuse/deescalate the situation? In the end this looks more like a misunderstandig of 'tone'.
well, split the cost in three and use cake's per IP fairness... and you solved most of the challenge. Or read up on how to implement your own hierarchical traffic shaper tree to implement the strict sharing regime you described.
Well have you tried nft-qos already and found it lacking? Or if you want ISP-grade configuration capabilities search for libreqos on github... (which requires beefy hardware to run on and is not something to run on OpenWrt....
That stock OpenWrt does not offer something like this out of the box just seems to imply that most users get by without such a feature...
So, do you want to limit the bandwidth, per device (as you stated on your first post)? or per person (as you stated on your last post)? Because (I guess) one person might have several devices (a laptop plus a phone, for example), and a device might have several owners (a TV in a common area, for example). Guests pose another challenge: whose bandwidth do they consume? How easily do you need to enroll their devices?
Depending on your answer, we will go down one rabbit hole or another.