The first thing you need to know is that the OpenWRT project is basically dead and that 95+% of the developers went over to the LEDE Project. However, LEDE has not yet published a stable release yet. You can get nightly builds that are in pretty good shape though.
I would highly recommend an ipq806x-based system, if you can afford it. Almost always matched with qca9880 radios. These are modern 802.11ac wave2 systems.
ipq806x is a Qualcomm-Atheros SoC. Go to wikidevi for specifications on the chips and all of the devices I mention below.
Check camelcamelcamel for recent pricing info if buying in the USA.
The list would be:
Linksys EA8500
TP-Link Archer C2600 (Not recommended due to TP-Link going anti-OSS. Modern versions require signed firmware and other DRM junk)
Trendnet TEW827DRU (Not yet accepted into LEDE, but could be any day now)
Netgear R7800 (Has a slightly faster CPU, but more expensive)
Netgear R7500v2 (Avoid the V1)
ZyXEL NBG6817 (Has the same slightly faster CPU as the R7800, but it's storage flash is goofy and I'm not 100% sure it's fully working. Ask the lede-dev mailing list first.)
The top issue that all of these devices have is that the 802.11 radio LEDs don't work yet because the driver is missing support for it. However, if you can live without blinking lights, these models are the way to go. This feature will almost certainly get fixed in the future.
I would tell you to go with the Linksys EA8500 if price/value is your concern. Otherwise the Netgear R7800 has a very active dev and probably has the best support. The ZyXEL NBG6817 looks really interesting to me, but I don't have one yet.
If $140-$200 USD is too much for you, look to some older 802.11ac devices. Like I said above, avoid TP-Link as they have started locking down their devices by removing serial ports and requiring signed firmware/DRM etc.
Your list here in comments is pretty good, though I'd avoid the TP-Link unless you can get one that is older (before TP-Link became anti-OSS.)
Calling the OpenWRT project dead sounds like hyperbole to me. As I write this, both OpenWRT and its package repository have had commits in the last day. That doesn't sound like a dead project to me.
I get that there is some strife, but calling OpenWRT dead would appear to be very very premature. FWIW I build OpenWRT from source and for the past few months have not been using CVS. I don't know the story behind it, but it looks like all development has moved to Github.
First off, OpenWRT as a project was very sick before the split. There was development going on with core stuff and bug fixes, but a lot of things were not getting done: Documentation, user-oriented outreach (forum, end-user help, new-dev help, etc), security response was a joke, and other many other issues. Some of the old core devs went AWOL and were really hard to reach and rarely made new commits or worked on the project, but they still had the keys to servers, DNS, and similar stuff.
Then one of the big dev servers over in Europe started going offline, and nobody knew why. Infrastructure like the user-forums, DNS, the bug tracker, and git/svn all would go down for HOURS or DAYS. People tried to reach out to the missing devs and got no reply... for DAYS.
That really kicked things off. People already were not happy with how the current set of project maintainers were running things. There were too few people with the keys to the kingdom and then when stuff broke, nobody was around to fix it.
The old devs are basically holding the name of the project hostage and preventing future progress.
So people forked and that's that.
The new LEDE Project has a lot of problems. I'm not a fan of how they have failed to scope out the project properly, but I think it's a better situation than the old OpenWRT project, which was dying a slow death. At least the new project has a chance at life instead of being smothered by the old devs who refused to share the keys to the kingdom with the active developers.
Sure, having competing standards isn't always a good thing but we're not talking about a standard here. It's a new router firmware distribution and I think having more of them is a good thing as it gives people more choice.
The major problem that I do see here is that LEDE has been forked from OpenWRT, but has nowhere near the level of device support that OpenWRT has. I have a TP-Link WDR4300 which I bought specifically because it was one of the routers which are highly compatible with OpenWRT. However, LEDE doesn't list this router amongst its supported devices (though there are newer WDRXXXX devices listed).
> However, LEDE doesn't list this router amongst its supported devices (though there are newer WDRXXXX devices listed).
What list are you looking at?
LEDE does builds for the WDR4300, and given that it's an ath9k router, it is and will continue to be very well supported. The only caveat is that I don't know whether they've added the necessary signature to allow the firmware to be accepted as a valid upgrade by recent TP-Link stock firmware, but once any third-party firmware is installed, there's definitely no barrier to upgrading to LEDE.
Hey, thanks for the great advice - here[0] is a comparison chart on Up/Downlink profiles and others from SmallNetBuilder for the first six models listed.
Do you have any real-life comparison of range/stability (and with what load on them?) on those models as from experience I know that raw comparison data on Wi-fi differs from reality.
Both this and some other Trendnet models I checked are not Wi-Fi Certified - has that been an issue for you?
As I'm in Europe the issue with TP-Link locking down firmware because of the FCC ruling [0][1] won't be a factor and I'm still not sure are they or TP-Link at fault?
On the contrary, I use the Archer C7 as my main router and in my experience, it is by far the best router I've ever had (having previously used FritzBox, Speedport, Technicolor, Netgear etc.) Here's a quick rundown of what happened: The part of the city I live in was upgraded to fibre for free (Fiber 200, with 180 down and 90 up guaranteed) and I needed a router that could handle that. The way Fiber 200 works is pretty simple: The provided optical end terminal converts the incoming fibre line to Cat5e. It then creates a VLAN, with separate IDs for TV, VoIP and Internet. So you need a router with unbounded VLAN ID config. Of course you can pay the $150 for the ISP provided one. But I didn't need VoIP, so the Archer C7 seemed perfect. Anyway:
Received C7 with old world-wide firmware. Immediately upgraded to the latest firmware, not realizing that this new version is region locked (no problem so far though). Tried to configure the VLAN for my Internet, realize that my ISP uses non-standard VLAN IDs which exceed the limit of the firmware. Call TP-Link (Germany). I was very surprised to speak to an actual engineer immediately. He knew the problem and recompiled the current beta firmware (with support for higher VLAN IDs) and sent it to me
Now we run into the first problem: You can't flash beta firmwares after upgrading to the region-locked one.
Call again - another very helpful engineer. He recommended flashing OpenWRT in recovery mode (FTP upload) and then flash the beta firmware from there. Flashed LEDE instead of OpenWRT (because I had it on hand), without any problems. Now I don't have any use for OpenWRT or LEDE. The functionality of the standard C7 firmware is more than enough for me. So I flashed the beta and everything worked.
Here are a few other points about this router:
- I pushed the internal switch to maximum capacity multiple times, even using unreasonable packet rates. The C7 doesn't even break a sweat.
- I use the built-in media server to stream music to my Dumb-TV. Works great.
- At one time, a friend brought his laptop over, which turned out to be infected. The C7 banned his device immediately, after it exceeded the DDoS protection limits I set.
- I never had any downtime on this router.
- WiFi performance is exceptional (up to 450M).
In conclusion: I would recommend the Archer over any router in this price range any day. The painless flashing of OpenWRT and LEDE makes it even more awesome. Add to that the awesome and competent support experience. And by the way: You receive an actual paper manual on how to request the GPL'd code that's used in the stock firmware.
How is a potential user supposed to figure this out?
If I go to what LEDE calls "ideal hardware for LEDE", at https://www.lede-project.org/toh/views/toh_available_864, none of the above routers are listed (with the possible exception of the Netgear R7500, but there is no mention of v1/v2).
I am very sympathetic to your comment. The old OpenWRT and new LEDE did a really poor job of end-user interaction. There wasn't a list of recomended models which was kept updated. The regular documentation was often out of date. The forum was poorly administrated. Project goals were unknown. Lots of back-room dealing with little transparency of what project direction should be.
Hopefully LEDE will change this, but I'm not very optimistic. It's the same devs as before, and they didn't like dealing with end-users before, so why start now? It took LEDE over six months to set up and end-user forum, and they decided not to make a lede-user mailing list at all. Not very end-user friendly actions I'd say.
Note that I don't interact with DD-WRT, so that's why I don't make comments about DD-WRT.
I would highly recommend an ipq806x-based system, if you can afford it. Almost always matched with qca9880 radios. These are modern 802.11ac wave2 systems.
ipq806x is a Qualcomm-Atheros SoC. Go to wikidevi for specifications on the chips and all of the devices I mention below.
Check camelcamelcamel for recent pricing info if buying in the USA.
The list would be:
Linksys EA8500
TP-Link Archer C2600 (Not recommended due to TP-Link going anti-OSS. Modern versions require signed firmware and other DRM junk)
Trendnet TEW827DRU (Not yet accepted into LEDE, but could be any day now)
Netgear R7800 (Has a slightly faster CPU, but more expensive)
Netgear R7500v2 (Avoid the V1)
ZyXEL NBG6817 (Has the same slightly faster CPU as the R7800, but it's storage flash is goofy and I'm not 100% sure it's fully working. Ask the lede-dev mailing list first.)
The top issue that all of these devices have is that the 802.11 radio LEDs don't work yet because the driver is missing support for it. However, if you can live without blinking lights, these models are the way to go. This feature will almost certainly get fixed in the future.
I would tell you to go with the Linksys EA8500 if price/value is your concern. Otherwise the Netgear R7800 has a very active dev and probably has the best support. The ZyXEL NBG6817 looks really interesting to me, but I don't have one yet.
If $140-$200 USD is too much for you, look to some older 802.11ac devices. Like I said above, avoid TP-Link as they have started locking down their devices by removing serial ports and requiring signed firmware/DRM etc.
Your list here in comments is pretty good, though I'd avoid the TP-Link unless you can get one that is older (before TP-Link became anti-OSS.)
Good luck