1 |
On Fri, 25 Aug 2017 17:46:01 +0200 Hanno Böck wrote: |
2 |
> On Wed, 23 Aug 2017 11:46:02 +0300 |
3 |
> Andrew Savchenko <bircoph@g.o> wrote: |
4 |
> |
5 |
> > Sigh... https also makes MITM attacks possible, especially if SSL |
6 |
> > or TLS < 1.2 is used or are allowed and protocol version downgrade |
7 |
> > attack may be performed. |
8 |
> |
9 |
> None of that is true. |
10 |
> |
11 |
> You're probably referring to attacks that were specific to certain |
12 |
> browser weaknesses, but they're irrelevant for this use case. |
13 |
|
14 |
Some attack are indeed implementation specific, but there are |
15 |
several which are design flaws, e.g.: |
16 |
|
17 |
1) BEAST attack[1]: TLS 1.0 is vulnerable regrardless of |
18 |
implementation (and all SSL versions). |
19 |
|
20 |
2) BREACH attack[2]: basically this is a side-channel attack for |
21 |
compressed traffic. All TLS versions are still vulnerable, the only |
22 |
practical mitigation is to disable compression. It can be argued if |
23 |
this is a vulnerability in TLS or TLS protocol has nothing to do |
24 |
with side channels, but if a protocol is vulnerable to a |
25 |
side-channel implementation-agnostic attack, it is considered by |
26 |
many as a protocol misdesign. |
27 |
|
28 |
Really SSL/TLS are very good examples of how crypto solutions should |
29 |
not be designed and implemented. |
30 |
|
31 |
[1] https://www.gracefulsecurity.com/what-is-beast/ |
32 |
[2] http://breachattack.com/ |
33 |
|
34 |
Best regards, |
35 |
Andrew Savchenko |