Are you an Ubuntu Linux user on either Ubuntu 20.10 (Groovy - unreleased, in testing) or 20.04 LTS (Focal - the current release)? The Ubuntu Foundations Team needs feedback on some possible major WiFi changes.
They're looking at replacing the currently widely used wpa_supplicant for iwd (iNet Wireless Daemon) a more modern upcoming solution which was written by Intel. According to Lukas Märdian, a software engineer for the Ubuntu Foundations Team it would give these benefits:
- Simplification of network management
- Faster network discovery
- Fast and reliable roaming
- Using less system resources
- Using features offered by the Linux kernel
- Support for enterprise security methods like EAP
- Support for kernel asymmetric key rings and trusted platform modules (TPM)
- Support for multiple clients
In the most recent updates to iwd it's apparently reaching feature parity with wpa_supplicant but it needs more testing. Something as major as messing with WiFI can have a lot of unintended side-effects, so getting plenty of testing done on a possible switch over is a good idea. From watching your favourite streaming services, to downloading and playing games online which is obviously important to us here.
Want to help? Find out how to test it from this linked post.
wpa_supplicant works with EAP, at least for eduroam(the worldwide federated university WiFi network). At least when used through network manager, unless Ubuntu has something hardwired to take care of that edge caseEAP is not an edge case. EAP is important for WPS too. So any wireless network controller that does not support authentication through EAP should not be used.
What I do miss though is WPS-push-button.
WPS is probably the only technique that securely can give any client a random 256 bit password.
<RANT>Okay, now we get tons of people that don't know WPS saying how insecure it is. It is not. There is only one variant of WPS that's insecure and that's the let's print a code on the device. Which goes directly against security. And most access points do not give a client a random 256 bit password, they give a shared password.
So WPS is absolutely more secure than a PSK when combined with openwrt.</RANT>
Anyway: https://wiki.archlinux.org/index.php/Iwd#Connect_to_a_network_using_WPS/WSC
it works it seems..
And eap for eduroam: https://wiki.archlinux.org/index.php/Iwd#EAP-PWD
Would switching to iwd change my life in any significant way? Or is it just small improvements in the background that make sense and a difference when being rolled out to the masses but on an individual level not so much?I would say, if it ain't broke, don't fix it in this case, but if you ever find yourself configuring wpa_supplicant manually, switch and don't look back, I swear that thing would just periodically break to hassle me for no apparent reason, if not upon further investigation, I'd find it was really due to... [insert obscure reason here], and really haven't had to give it a second thought since switching to iwd which is really how it should be.
I'm just curious, I'm not on Ubuntu anyway.
That said, I'm just using PSK at this point and I do know the equally frustrating hassle of trying to configure some crazy EAP thing in config files so, in that case you'd maybe be better off not switching until it's fully supported by NetworkManager or whatever config tool you're using supports it as well. Probably the question Ubuntu will be answering here, does it support most user configs well enough and that idk, but all else equal it definitely seems to be an improvement other ways I can tell.
Plays nicer with systemd-networkd and dbus for sure, or no other config service at all since it can handle basic DHCP, just maybe not fancy things like "use ISP's DNS but enforce DNSSEC so at least i know they aren't evilly changing records"
It would be one thing if this Lukas Märdian were asking users to test out iwd because it was being considered as a replacement for wpa_supplicant. Fine, let's give it a shot. But that's not what Lukas is asking at all. Lukas is asking users to see if iwd is "ready for prime time," that is, if it's ready to become Ubuntu's new default.
Why is this even happening? Iwd isn't better than wpa_supplicant. There's no promise that it ever will be better. There's just some vague hand-waving and talk of "potential."
Does Canonical make these decisions using a Magic 8-ball?
Question: "O Magic 8-ball, should we make our own mobile phone from scratch?"
*shake, shake, shake*
Answer: "Signs point to yes."
Is iwd even open source?
I love Ubuntu, but gawds it makes me crazy sometimes.
Last edited by Nanobang on 14 August 2020 at 2:11 pm UTC
Is iwd even open source?
Of course it is. And it has been in your repositories for quite some time already. It's just not installed as default. You can find more about here for example
https://git.kernel.org/pub/scm/network/wireless/iwd.git/about/
Is iwd even open source?It wouldn't have a damn git repo on kernel.org if it wasn't open source, now would it? Took like ten seconds to check.
I love Ubuntu, but gawds it makes me crazy sometimes.
The reason there hasn't been much debate about this is because iwd is simply better than wpa_supplicant, and the only question is if it's ready for prime time. That's why Canonical is looking for feedback. Note that this isn't a Canonical project. They've had their NIH moments, but this isn't one of them.
Check out the iwd wiki for more info unless you'd rather just shout at Canonical to get off your lawn. ;)
Is iwd even open source?
I love Ubuntu, but gawds it makes me crazy sometimes.
It wouldn't have a damn git repo on kernel.org if it wasn't open source, now would it?
Frankly, I don't know.
I did follow links there from Liam's article; I know nothing about kernel.org --- and really, still don't; though I looked at the page you linked. I didn't see anything about what license iwd was or wasn't being released under, I made no decisions one way or another. What I did see on that page, at the very top was:
Copyright (C) 2013-2019 Intel Corporation. All rights reserved.
iwd is simply better than wpa_supplicant
Thanks for telling me that, I'll just take your word for that then, shall I? Because Lukas Märdian's post quite clearly states that iwd hasn't even fully reached feature parity with wpa_supplicant, that iwd merely offers "potential benefits" over wpa_supplicant. I didn't say iwd isn't better, just that that's not what's being stated.
Note that this isn't a Canonical project.
Yeah, I got that it's Intel's thing, but the decision to use it is being made by the same Canonical cum Ubuntu crew that's given us all those previous NIH moments you mentioned.
Last edited by Nanobang on 16 August 2020 at 2:54 pm UTC
See more from me