[sulandor]

what is the rationale behind this effort? are wifi-ap's moving towards x86?

ime a usb wifi stick in a desktop computer will work as wifi-ap, but is somewhat janky because of the metal case (which is needed because em-interference from bus-clocks) and the wifi hardware having suboptimal provisions for ap-mode.

UPDATE:

"because we can" (was a stupid question)

no hard feelings; last used tomato ~20y ago on a wrt54gl

[ssl-3]

Tomato is just a Linux-ey system that routes packets and which features an approachable GUI, and an access point is just a system that bridges wireless networks to wired networks.

These functions are normally wrapped up in one box in the consumer space, but they're still very different functions.

And maybe I'm not doing it right, but I myself haven't used a combined router+wireless box in a fair number of years at home or at the shop.

I keep the wired networking+routing back end in one spot where it makes sense, and I keep the wireless access points where they make sense to provide good coverage where I need it.

My router just routes, and my wireless access points just provide wireless access.

---

So to answer your first question directly: This system lets people use the friendly Tomato system on any old (or new, or whatever) x86 hardware they have. It brings it out of the world of hacks[0] on cheap low-performance embedded Wal-Mart routers and lets a person use it for routing on a much more performant machine.

[0]: Not that those things aren't fun. I still have the first standalone router I ever bought -- a Linksys WRT54GS, with Tomato installed, and with an SD card hacked in using a card-edge connector from an old floppy drive cable for expanded storage.

[leptons]

I've been using DD-WRT x86 for a few years, so it's very similar to what Tomato x86 is doing. When I upgraded to 1gbit internet, my old high-end Netgear router that I paid hundreds of dollars for couldn't keep up. And I knew 2Gbit internet was coming, and now they're offering up to 7gbit fiber speeds.

New routers with faster CPUs to keep up with the increasing bandwidths are costing about $700. No way am I spending that kind of cash for a router, even if it has the latest Wifi.

Instead I bought a cheap $50 Dell from ebay with a quad-core i5 CPU, I installed DD-WRT x86 on it, I put in a cheap 4 x 1gbit ethernet card, and I bought a cheap refurb Wifi 6e router and use that only for the wireless functions. All-in it's about $200. Now I can keep the same main x86 router hardware and I can keep upgrading my internet speeds as well as upgrade the wifi externally whenever I find it necessary (and I can find a cheap wifi radio).

The DD-WRT maintainers were also very helpful (and in a timely way) when I requested they add a 2.5gbit NIC to the DD-WRT drivers, so now I have upgraded to a 2.5gbit network. And if I want to put in a 10gbit NIC someday, maybe they'll help out again. I know, I should move to something "more modern" than DD-WRT, but it suits my needs well for now, and someday I'll probably be moving to a different x86 based router software should DD-WRT not be able to keep up with my needs.

[colechristensen]

Use case would normally be a small form factor x86_64 machine which can be price competitive with high end AP hardware and you don't have to play the game of finding compatible hardware (squatting in the aisle in a walmart looking at the hardware revisions of all of the APs they have in stock finding out that they have the AP you were looking for but the hardware revision was incompatible)

Being able to buy a "normal" computer and install tomato on it is a more attractive prospect for a lot of reasons.

[tencentshill]

It's quite expensive to get a 10GBe-capable home router and switch, but this now allows you to use any old x86 PC and a 10GBe networking card to create one of your own. That's what I'm going to try.

[sulandor]

10gig software switch is easy, 10gig software router medium, 10gig software dpi challange

it's a nice exercise but the power and space requirements in relation to the performance will generally not be favorable unless you are severely constrained in up-front budget