Talk:Snowflake (software)

Traffic in Iran
Tor's usage graphs filtered for Iranian origin show Meek traffic rising drastically, then crashing and being partly replaced by Snowflake traffic, which rapidly crashes. From the bug reports, it seems the devs accidentally took down the server relaying Meek, and lost the private encryption keys; it's now back up with new keys. The crash in Snowflake use was the Iranian government sucessfully doing TLS handshake fingerprinting; Snowflake used a very weird, distinctive order in which to list part of the handshake. This is also fixed in the new version. Use in Iran has not recovered; perhaps Snowflaake hosts don't have the new version, and Iranians don't have the new version of Tor Browser with the new Meek public keys. I have not been able to find any third-party coverage of this, odd given all the news on Iran; if anyone reading this knows of such sources, please mention or add them! HLHJ (talk) 16:22, 8 November 2022 (UTC)

Snowflake is not Tor
It's merely a tool to access Tor where it is blocked. So I think we need to remove the second and the third paragraphs, and make edits across the page. In theory Snowflake can be used with no conjunction with Tor at all. Describing Tor when we're supposed to talk about Snowflake is confusing. WofWca (talk) 18:36, 14 November 2022 (UTC)


 * WofWca, do you have any sources describing that use? I got the impression it was just a Tor pluggable transport. Thanks for the info! HLHJ (talk) 22:31, 27 November 2022 (UTC)
 * Sorry, I didn't mean that it is used right now for anything other than accessing Tor. I'm saying that it takes minor adjustments to use it as a general censorship circumvention tool. So what I'm saying is that Tor can exist without Snowflake and Snowflake can exist without Tor. WofWca (talk) 19:53, 4 December 2022 (UTC)
 * See the "Future work" the draft of the research paper:
 * > A natural extension of Snowflake would be to have it access systems other than Tor—ordinary VPNs, for example WofWca (talk) 19:08, 5 October 2023 (UTC)

A grammatical issue with one of the sentences in the article
Hi.

While reading the article, I came across this sentence in the Function section:

"The traffic looks like ordinary peer-to-peer traffic, such as is used by many videoconferencing apps."

I understand its meaning, but think it's grammatically wrong; IMO, the phrase "such as" must be followed by a noun, not the verb be, or something similar to a passive sentence!

I'm really curious, and I would appreciate anybody who can give me more info. Javad78905 (talk) 02:24, 10 February 2023 (UTC)


 * See 1151046440, do you think it's correct now? Shinohara Chihiro (talk) 15:39, 21 April 2023 (UTC)
 * Yes, you actually corrected it with your edit. Now it looks like:
 * "The traffic looks like ordinary peer-to-peer traffic, such as that used by many videoconferencing apps."
 * After "such as", we have the word "that", which functions as a noun. Then it's followed by a relative clause (shown in bold). The relative clause might begin like "which is used by ...", but "which is" is omitted from the beginning. So the grammer is now OK. Javad78905 (talk) 14:43, 27 April 2023 (UTC)

A Persian version of the Snowflake article
Hello guys!

I've just published this article (Snowflake) in Persian, and I'm really excited about that! Translated into Persian with love by Javad Bayat (me), a student majoring in computer science at Isfahan's Mohajer university.

Sincerely, this is the very first Wikipedia page I've ever translated fluently and published. So, if you have any suggestions/issues about the translated article, please don't hesitate to let me know, either by emailing me at javad78905@gmail.com, or by contacting me via Telegram at @javad0bayat. Javad78905 (talk) 09:24, 20 March 2023 (UTC)


 * Why dont you enclose any wikipedia page links to your article?
 * There's seems to have 2 draft pages on fawiki: fa:کاربر:Javad78905/دانه برف (نرم‌افزار) and fa:پیش‌نویس:دانه برف (نرم‌افزار). Shinohara Chihiro (talk) 15:45, 21 April 2023 (UTC)

A research paper from Snowflake authors
Snowflake authors have published a draft of a research paper titled "Snowflake, a censorship circumvention system using temporary WebRTC proxies". It can serve as a great source. WofWca (talk) 19:06, 5 October 2023 (UTC)


 * WofWca, that's really useful, it gives me a much better idea of how it works. I'll have a go at fixing the article to be less Tor-specific. The diagram in it is also clearer to me than the one we have in the article; would it be possible to upload it to Wikimeida Commons so we can use it here? If the article is published under CC-BY or CC-BY-SA or CC0, anyone could upload it under that licence, but if it isn't licensed the artist (assuming they hold the copyright) would need to be the one to upload it. HLHJ (talk) 03:08, 10 March 2024 (UTC)
 * BTW they updated the paper where they also updated the picture with STUN servers and indirect rendezvous, you can see the freshest build here, and the new diagram can be found here.
 * I'm not sure about the license though. I'll ask them on that forum thread.
 * I've also found another diagram that is under CC BY, arguably it's more strict and thorough, though doesn't include STUN servers. WofWca (talk) 07:56, 10 March 2024 (UTC)
 * Thank you for asking about the image, . I think that what makes that diagram look clearer to me is the way it puts the broker server off to the side and keeps the main traffic path in a straight line. This isn't a critique of the academic paper, but an encyclopedia article has a very different audience: almost anyone who might want to know what this Snowflake thing is. So that simple, top-level-overview diagram is ideal to start with. The Bamsoftware-hosted version seems to me not as encyclopedic as the older version you linked to in the first link in this section; the paragraph references to the paper won't work in this Wikipedia article, and the STUN servers are a complication which isn't essential to getting the basic idea. The MDPI article is nice to know about, and appropriately-licensed, but none of the images seem quite high-level enough (unless someone adds a LOT of detail to the article ). I have been re-drafting the text, and will post when I have some time to make sure it's solid. HLHJ (talk) 00:38, 14 March 2024 (UTC)
 * > different audience
 * I guess I agree. Though those who want to go deeper will definitely wonder about STUN, then they'll need a more detailed diagram. WofWca (talk) 19:23, 16 April 2024 (UTC)
 * On the minor-comments front, I'm surprised that alphabetical order of self-disclosure lists-of-capabilities isn't an RfC. HLHJ (talk) 02:22, 14 March 2024 (UTC)
 * Sorry, what are you referring to? WofWca (talk) 19:24, 16 April 2024 (UTC)
 * So, they released the diagram to public domain!
 * Plus, the paper has been finalized! WofWca (talk) 19:19, 16 April 2024 (UTC)