求助client hello标准 / help with client hello standard

[louiesun]

见此https://github.com/URenko/Accesser/discussions/201#discussion-7275050

See here: https://github.com/URenko/Accesser/discussions/201#discussion-7275050

[wkrp]

@louiesun Do you have a specific question to ask? The linked discussion seems to be about how fragmentation of TLS handshakes sometimes succeeds against the GFW, despite the GFW's ability to reassemble TCP streams. But I do not understand what, specifically, you are asking about.

[louiesun]

@louiesun Do you have a specific question to ask? The linked discussion seems to be about how fragmentation of TLS handshakes sometimes succeeds against the GFW, despite the GFW's ability to reassemble TCP streams. But I do not understand what, specifically, you are asking about.

I used to find that tcp frag sometimes bypass gfw and i wanted to find out why.

but now i find that tls frag can bypass gfw, so the question is not necessary now and we can close it.

[louiesun]

So notice! There are two kinds of frag. one simply frag the clinet hello to several pieces on tcp layer. but infact, tls layer supports frag, it make fata into pieces and add a head to each of them.

In china, gfw assigns the tcp frag (though it sometimes mistakes) and rst, but it doesn't assign tls frag.

[wkrp]

Yes! You are correct. Here is a thread that touches on the difference between TCP and TLS fragmentation: #308.

This blog post explores an interesting idea: fragmenting TLS messages (especially Client Hello) over multiple TLS records. This is different from TCP segmentation, which has been studied in the past by, e.g., Winter & Lindskog 2012 (Section 5.2, brdgrd), Khattak et al. 2013, and Bock et al. 2021 (Section 4.1), and is implemented in tools including GoodbyeDPI. Rather, this research takes advantage of the fact that TLS messages (e.g. Client Hello) are carried in TLS records, and that one message may be fragmented over multiple records.