Jump to content

Talk:WireGuard

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
[edit]

Somewhat concerned about the number and proportion of cites that link to pages with requests for monetary donations to this project and feel it would be better if that ratio could be reduced. Thankyou. Djm-leighpark (talk) 08:09, 25 June 2018 (UTC)[reply]

They shouldn't be reduced, they should be removed entirely. I've removed the blog refs, but I don't think an archived URL that is needed to source article content should to be removed — you can't really make donations through a dead link right? I don't want to remove a ref that sources content without consensus. Seraphim System (talk) 08:20, 25 June 2018 (UTC)[reply]

Merge to mainline linux kernel

[edit]

Efforts seems ongoing to do this. Seems to have missed 4.18 in June 2018. Was tempted to add to article but halfway backed out. Couple of plausible information I found were at :

— Preceding unsigned comment added by Djm-leighpark (talkcontribs) 09:54, 25 June 2018 (UTC)[reply]

Supported Versions Query

[edit]

FOSS normally doesn't 'support' AFAIK ... though a commercial company doing a supply based on their codebase might. Do we need to cite the OS list as Android, BSD aint Linux ... and the developers seems to be cautioning against Windows client currently in the install list.Djm-leighpark (talk) 23:15, 25 June 2018 (UTC)[reply]

Concerns over Design aimsFeatures section

[edit]

I have a concerns over the the content added to the new design aims section, primarily I feell as inadvertently it might be judged promotional to the website supplying the associated reference. This is due to use a bare url; emphasised use of the website name twice; and because this has also been added as the first (most prominent) section of the article. I feel there is also some possibility the 'design aims' content of the review may not be supported by the reference .. but I'm open to that. I see no sign of the user having done this sort of promotion elsewhere, and the reference looks quite useful in terms of citation for article development. So every chance is is all good faith; and may have been done to specifically emphasise the provenance of the cites in the article. I have already changed the bare url into a cite and checked and recorded the content as already being available in wayback and not subject to linkrot. I will probably remove the website name from the article as its available in the cite as that help alleviate any promotional issues. I may move the section, or part of it, to a section named 'Reception' at the end of the article and I may put cn on individual sentences ... I am currently thinking about these things.Djm-leighpark (talk) 14:53, 18 October 2018 (UTC)[reply]

  • I remain unsure about many comments above so currently leaving as is, though on careful review I have decided most if not all of the section contents are addressed in the source. I have however changed the section title to Features as I feel this is more appropriate. There is little gain in making this other than the first paragraph at this stage. I remain unsure about use of the website name twice in the section but at this point there are also some arguments for a straightforward removal ... so on that point I am currently inclined to leave there myself at the moment.Djm-leighpark (talk) 18:31, 18 October 2018 (UTC)[reply]

June 2019

[edit]

I observed Wireguard didn't make the Linux 5.2 March 2019 cut .... but in June 2019 I've come to notice an August 2018 Ars Technia review here [1] (With LT quotable), Another ArsTechnica review about the windows pre-alpha (https://arstechnica.com/information-technology/2019/06/taking-a-spin-on-wireguards-windows-pre-alpha/) and the attack surface document [2]. All possible to incorpoate to the article.Djm-leighpark (talk) 07:19, 28 June 2019 (UTC)[reply]

Still feels like a promo

[edit]

It's now live production code, so it's time for the article as a whole to focus on what it achieves, rather than what is still planned.

There are still far too many words that are synonyms of "intention" or "goal". Statements that include those words belong in a "rationale for development" section, not scattered throughout the whole article. (There may even be room for a "goals not achieved" section.)

And there are other hyperbole words that need weeding out:

  • In the introduction: better performance and more power -> what does "power" add here? suggestion: change to better throughput and quicker (re-)establishment (since the latter is in fact one of its aims).
  • In section "Reception": have relatively massive code bases of over 400,000 and 600,000 lines of code, respectively, which ... -> suggestion change to have a million lines of code between them, which ....

Martin Kealey (talk) 07:00, 7 September 2021 (UTC)[reply]

Remove citation-needed

[edit]

There is a "citation needed" on Since point-to-point is supported, other topologies can be made, but not on the same tunnel, which I believe is spurious as that's a sky-is-blue kind of blindingly obvious statement for anyone with any experience as a network engineer.

This kind of thing is hard to cite because once you know it, it's so obvious that you forget where you learned it. But if I had to guess, I'd recall that the first lecture of any course on networking starts with a description of the ISO & IETF layers, and their logical independence. From there it's a very short step to the observation that any layer-3 topology can be constructed on top of any layer-2 transport, including tunnelled point-to-point links.

Martin Kealey (talk) 07:17, 7 September 2021 (UTC)[reply]

Eye twitching

[edit]

There are numerous references to IPsec as a VPN protocol. IPsec is *not* a VPN protocol, it's an encryption system built-upon IPv6 and backported to IPv4. It is often used as a layer underneath a simpler VPN protocol such as L2TP to provide encryption, but that's not the same thing at all.

The fact that the term IPsec is abused so often to refer to VPN protocols that use IPsec for encryption is actually making it harder to find information on IPsec itself. Most attempts to Google IPsec give you useless VPN configuration pages instead.

Please stop doing this. If you have a VPN protocol in mind, name it, not some technology it uses. Imagine if we started referring to all VPN protocols as "IP" or "Internet". --2601:584:300:B0A0:15B0:DB5B:31B1:62E6 (talk) 15:54, 13 November 2021 (UTC)[reply]

Unfair Comparative to OpenVPN

[edit]

I agree with the the "Still feels like a promo" topic above, especially considering this line: "This is unlike alternatives like OpenVPN because of the many disadvantages of TCP-over-TCP routing."

OpenVPN can operate in TCP or UDP mode, depending on the configuration. You can have a TUN driver implementation of OpenVPN configured as TCP or UDP. See https://openvpn.net/faq/why-does-openvpn-use-udp-and-tcp/ - "The OpenVPN protocol itself functions best over just the UDP protocol. And by default the connection profiles that you can download from the Access Server are preprogrammed to always first try UDP, and if that fails, then try TCP"

Therefore, this article has an unfair comparative to OpenVPN with its statement, further appearing to be promotional. Lesterclayton (talk) 06:56, 8 August 2022 (UTC)[reply]

Adoption in commercial products

[edit]

Just to point that CATO Client's Windows virtual adapter, named "Cato Networks Tunnel", under the Manufacturer property reports "WireGuard LLC". Does anyone have better informations, so that this could eventually be added to the "Adoption" section of the main article? Regards

IlGino (talk) 06:38, 22 May 2024 (UTC)[reply]