Hi.
Regarding H/W OpenWRT one will be the historical device in the human history. It's obvious as the small step. But... if compare with commercial devices, no advantages in design, performance vs price, etc...as you already aware of.
Since I have no information, I suggest the next device.
Frist, for consumers,
make nas-ready device which have below hardware outfit but OpenWRT.
RAID support (2 by 2(for two mirror-RAID) or single 4 slot-raid)
Not only SATA interface, but NVMe PCIe variant.
This will be used for home or SOHO purpose as storage or torrent.
Important thing is the OpenWRT preflash.
You can make all prebuilt component as a set, or sell it as component and release 3D stl case model on thingiverse. So that can deliver around the globe with less weight and size. Check out how the Ploopy shop guys doing.
Second, for business,
make the router that metal sheet outfit(like OpenWRT one) as a software defined router
which this guys doing(SDR by OpenWRT).
As a reference H/W board design, you guys could plan to get some royalty for the community operation. Or sell openwrt-self include corporation support package like redhat business doing.
What I'm personally interesting is the first one which really need. Quality is the matter.
At minimum(4 modes) : No raid(running seperatly) / Raid 1 / raid 0 / large mode(aka JBOD)
(which is defacto firmware raid mode by hdd enclosure raid)
Full spec : Support RAID 0/1/3/5/10/JBOD/Single Mode. refer full spec
Plus, as I mention that '2 by 2' mode means, two separated raid1 mode for 4 bay model.
or raid 1 and two single drive mode. What I'm trying to say is a user can be choose ANY combination of drive mode. That must be strong point of OpenWRT base device.
For #2,
Right. 1U, 2U kind physical specification need to be followed.
But the 'L-shape mount parts with screw' can be separated to be used table type for safety.
some of model take this way. Not mono body. No need to.
Even though 'business model', not 'Corporate' grade niether 'Enterprise' grade for the time being. Need to aiming small office/home office whcih less 50~100 people throughput capacity.
Need to consider the RJ45 port count (8 ? 16? 24? varients, or even module base port extender?(like traditional blade server)), and CPU performance, important thing is expernal interfaces for USBs, thunderbolt(storage and so on), or any famouse NFS interfaces in datacenter industry.
I would argue that, while one perfectly can (and I actually do) use OpenWrt on a NAS, it is decidedly not its primary focus and creating NAS hardware is entirely out of its scope. Open-source NAS systems do exist, you would probably have an easier time looking at an x86 enclosure with TrueNAS, OpenMediaVault, Unraid or similar.
Small business/big home maybe if they have some IT-neerd in-house responsible for the in house network.
But still, now we talk about all their actual saved data and not routed data which any router can keep on doing.
Why don’t they instead buy a purpose designed NAS from Synology or Q-nap with warranties that the raid system actually work?
Realize what the suing sums would be on OpenWrt if they loose all data because of a faulty raid system?
A NAS from the real NAS manufacturers would also pretty much be guaranteed to be cheaper than a more expensive small manufacturing volume non warranty OpenWrt design without future proof?
That is the fact of OpenWrt, the router part is pretty future proof, but all these side project are not future proof. They are only as future proof as the one single developer being personally interested for the time being.
To be future proof for OpenWrt also requires that the open source drivers for the hardware actually keep on functioning in the kernel and to this date OpenWrt has no access to closed source drivers so OpenWrt can’t use state of the art dedicated components either since the drivers for those are probably closed source since they are worth money.
But on that lower business level OpenWrt would also be required to explain why an expensive NAS is better than some cheap cloud solution?
If we then move up to the next business level, I can guarantee not a single corporate or enterprise size business will ever look at OpenWrt as a IT-infrastructure provider. It will simply never ever happen.
That business level never run their own IT equipment to begin with, instead they always sign big multi million/billion $ service contracts directly with big IT suppliers like Cisco, D-Link, TP-Link or similar that can deliver a complete big IT solution with guarantees, warranties and 24/7 support and service obligations.
The issue with a self flying taxi is that it would be unsafe to let customers flash custom firmware on it. And a lot of the OpenWrt ethos is about taking ownership and customising your own products.
As such, the OpenWrt Four will clearly be a self flying personal transport solution instead. Such as a flying electric bicycle or scooter. I predict interplanetary travel is unlikely until OpenWrt 9 or 10.
