Poll - Interest in an OpenWrt purpose built router

Some of the manufacturers take a modular approach. ZBT for example, stuffs both MT7628KN and Realtek based boards into the same case. I think we can find a manufacturer with an already RF tested board and registered with the regulatory agencies. OpenWrt supplies basic firmware to the manufacturer who does the initial flash.

The manufacturers web sites state they will warrant the hardware but I would also get that on paper.

Kind of on my way out the door so in a hurry, sorry if this idea has already been brought up and beaten to death but when I see this I think of something that starts here:

https://openwrt.org/inbox/toh/pine64/pine64_rockpro64_v2.1

Has anyone considered contacting https://www.gl-inet.com/ ? They seem like a small outfit. I've bought 3 of their routers over the years and they run a lightly modified version of OpenWRT. The routers have generally been underpowered (though fine for me with only 30 Mbps of WISP bandwidth) but if there is interest I'd bet they could be convinced to build something more substantial as they already seem headed in this direction.

3 Likes

It would be nice to have atleast 8 ports with POE support.

2 Likes

Their business model actually uses kickstarter so I guess this project would be competing, although only on one device initially, I do not see anything on their site that they contribute/support OpenWrt in anyway.

One advantage, it would take care of distribution/shipping. So we could present what were looking for to them with a promise of 500+ orders. See what it would cost with a 20% OpenWrt donation/conference support etc. Thoughts on this?

Learning from the past is a good idea. This is what I've been dealing with when it comes to "custom hardware for routers" :slight_smile:

https://www.kickstarter.com/projects/itus/iguardian-the-home-internet-security-system

https://www.pcmag.com/news/why-hardware-start-ups-need-to-be-careful-when-crowdfunding

3 Likes

@Grommish
Thanks. On the minus side, kickstarter sites take a 5% cut and the percentage of failed projects is high.

On the plus side, what I envision is very simple, essentially taking an existing product and putting OpenWrt firmware/logo on it.

In parallel to this thread, I'm looking at porting OpenWrt to a Trendnet TEW-827DRU V2 (mt7621, mt7615E x2) and have been spending more time trying to get access to the device (bootdelay = 0) than actually doing any OpenWrt development. Supporters of the project can go directly to the sdk without having to hack though locked bootloaders, cameo signatures, etc.

At the end of the day, I'm going to close the poll, after 14 days and see if I can get a rough idea of the number of views. From this sampling, about 82% would be interested. Will contact the OpenWrt branding email and post to the OpenWrt-devel list.

1 Like

Just so you are aware, that Kickstarter did work, they did deliver (I've got two of them).. they still failed after the fact.

I'm certainly not being negative towards the idea, but realize that it tends to be more difficult than just that. Getting the money doesn't seem to be the hard part :slight_smile:

Besides, if a ODM is created, there isn't a reason OpenWrt's implementation for that device couldn't receive legal clearance to include things like the proprietary drivers that seem to be needed to get the most out of the hardware.

1 Like

Yeah... that ITUS device highlights IMHO what is the first of key of two attractions/values to this concept...

  • Tinkerer / Advanced Hardware featureset ( not found off the shelf )

( the second being official sanction/code optimization/support for ~all hardware components )

The trick is finding a balance between;

  • unique / performant / contemporary
    and
  • generally suitable / comparable or better than whats on the market ( i.e. multi-Gbps ) whilst maintaining interoperability, cost effectiveness etc.

Simply rebranding or "ordering" a rebadged commercial offering has no point of differentiation.

I'd love a switch on the front that let you choose "personalities/dual-triple boot"... ( not in the exact way that they did it - but those things can be fine tuned )

Dual-boot is indeed a great feature, but fortunately it's increasingly found on mid- to top-level devices.

1 Like

I'm not sure this is doable, though I'l like to see such a device. Whose needs would it fulfill? A decent ethernet-only router can already be found, as well as those with ac wifi.
I'd throw all of my money at a fully OpenWrt-supported device with a modern 4-core ARM CPU, at least 5 gigabit ethernet ports, VDSL+G.fast modem, SFP+ slot for future-proofing (for when I get fiber around 2090), a decent ac wifi and 2 FXS ports. Anything less than that is already available on the market. Anyone willing to make one with my requirements? :slight_smile:

Some follow up on this. contact@openwrt.org is how one inquires about utilizing the OpenWrt brand and logo. I did not receive any response.

I posted at openwrt-devel and received this:

> request.
> Reply-To:
> Organization:
>
> I sent an email to contact@openwrt.org, subject line Trademark Use
> Request, inquiring about crowd funding an OpenWrt router.  I never
> received a reply.  The email gave the background, links to forum
> discussion and a poll and outlined the basic structure.  The contents:

Probably nobody with access to this mailbox was interested in your idea.

[...]

> Would the project be interested in exploring this further?"

Personally, I don't think so. 3y ago we (some of devs with commit access) were
discussing similar idea. But there just wasn't enough interest in it.

1 Like

OpenWRT seriously needs their own WRT3200ACM type router, powerful with open source drivers.

This is an R/F device, therefore regulatory certification like FCC or local equivalent is imperative for legal distribution of any custom designed solution. If not approved in a given local jurisdiction, legal distribution would be impossible and the first to launch a complaint to authorities would be the name brand vendors, therefore causing a ban/seizure of illegal hardware at the import level.

Unless OpenWRT is willing to undertake such expensive certification process cost, the custom design discussion must come to an end, period.

The only other viable solution would be pre-certified white box label of pre-existing/certified designs, bought in batch process. However, in my youth, I must have overseen the assembly of thousands of clone PC'S with parts manufactured by known Asian manufacturers. I can vouch the fact that quality/flakeyness of various parts/components varied very much depending on manufacturer, product and production run/batch or even revision for that matter. In the end, a very risky endeavour for OpenWrt.

In the end, it would probably be wise to just pick a half dozen commercially available products and make sure they are properly supported.

For example the Linksys EA8300 is supported, has ample RAM and Flash, decent SOC, tri-band and more importantly dual failover boot/firmware partitions. The R7800 seems to be a very popular device as well. The Rock Pi E and NanoPi R2S have descent SOC, RAM and unbrickable due to removable Flash, however if they were better supported, I am sure their popularity would skyrocket.

In the end, it's about KISS, Keep It Simple Stupid.

There are several devices out there, addressing different market segments/requirement, it would be wise to maybe figure out a short list of devices and just support them to death.

KISS... It's not about reinventing the wheel, it's about picking up a good set of wheels and having a good drive...

Just my two bit rant...

Edit: The MR8300 is basically same as EA8300, except more RAM, therefore supporting to death families of a short list of equivalent hardware, makes for more choice as well.

5 Likes

I guess the only solution would be to go to an existing manufacturer like TPLINK and let them make the hardware for you.

You still would have to pay for local regulatory certification, likely very expensive due to relatively small production run, cost per unit would be prohibitive.

I ll stick to my WRT3200ACM as long as I can, I really have no problem with it. I even got a matching Linksys SE4008 switch for it.

2 Likes

I'm not sure this is true. The device in this link is identical to the Newifi D2. The same device was marketed by Lenovo. The Newifi D2 presently is OpenWrt supported. Newifi used the Breed bootloader and provided user options for OpenWrt, Padavan, Asus and other firmware downloaded from their site.

The Newifi D2 is in the FCC database and it is highly likely that Shenhzen-EDUP built this for Newifi and Lenovo. The manufacturer, in the link, would only have to attest, in a letter sent to the FCC, that the hardware is identical. This nominal legal cost could be negotiated in the bulk purchase price - not sure what the FCC would charge to process the letter. OpenWrt could provide the firmware to the manufacturer for the initial flash.

I was commenting with respect to custom hardware, not relabeled models, in which case a precertified relabel would alreadly have some of the certification costs offset by previous production runs indeed.

Furthermore, if there were an OpenWrt labeled product, there’s an issue with product liability and insurance, I’ll skip the details but there would also be an ongoing yearly cost for Liability insurance polcies.

The details are important. If you are referring to a warranty, the vast majority of warranties are voided when you flash OpenWrt. Look at the Cudy AC1300 thread on the Cudy supplied OpenWrt software and look at the diclaimer on Cudy's website.

If you are talking about liability due to a device catching fire, that would be the hardware manufacture - the same who certified to the FCC.