SWEET32 : triple-DES should now be considered as “bad” as RC4. · Issue #387 · ssllabs/ssllabs-scan · GitHub
It looks like ghost from past has visited as again. From official problem disclosure web site  there is an info: "The Windows XP and Windows Server 2003 [...] they support only RC4, Triple-DES, DES, and RC2-40. [...] situations where the best available cipher is Triple-DES."
Two years ago there was accepted in Firefox that 3DES should is preferred to RC4 . Now we also have RC4 broken, AES-CBC broken with TLSv1.0  and now 3DES broken.
Impact is as following : "We performed a scan of the top 1 million servers as listed by Alexa using the cipherscan tool. We found that 86% of the servers that support TLS include Triple-DES as one of the supported ciphers. Moreover, using the ciphersuites sent by a modern client, 1.2% will actually use a Triple-DES based ciphersuite."
We can get into another debate what now? Should we disable 3DES on server? Most probably not if there are old clients accessing web server or like Mr. Rob Moss has written on this forum : "Where the cut-off point lies depends on the size and the expectations of your target market, and that unfortunately has nothing to do with security."
Also interesting what RobTho has written on bug tracker : "The final nail to IE@XP's coffin (reference to link 6)". If looking at  it is obvious IE on Windows XP the 3DES was until know the only secure protocol.
 Qualys SSL Labs - Projects / User Agent Capabilities: IE 8 / XP
Where you find problems is with websites which are written for internal corporate use, where the company involved lacks the resources to keep operating systems, web browsers and hardware up to date. These sites aren't used by normal web users - it's a specific set of users with a specific set of requirements. And even broken TLS is significantly harder to eavesdrop on than plaintext HTTP, so these ciphers (RC4 and 3DES) still have some use cases, even though we'd prefer everyone to be using AES-GCM or ChaCha/Poly. Even these modern cipher suites have potential problems with quantum cryptography, although not quite so bad as everyone believes, given current research.
I think the general advice on 3DES is quite simple: drop it. Those people with a genuine reason to continue to use 3DES will know that they have a genuine reason to continue to use 3DES, and they're a very specific bunch. Of the 860,000-odd servers supporting 3DES, I would suspect that there are fewer than 1,000 with a genuine reason to continue to do so.
Again (and again) we see that most affected installations are driven by client needs, not by security. Instead of showing the Red Card to Win/Lose XP installations, the admins will argue that "the market in China is still based on XP" or that "we cannot ignore those 3.8% clients still on XP". Today's business might be affected a little, yes, but tomorrow's business will benefit from consequent high security setups. And all those security threats "from the past" will just pass by.
Such people which still use XP and do not want or cannot change this OS, for whatever reason, can easily switch to Firefox, which still supports XP and modern ciphers. And i beleave they will try another browser if they find IE stopps working.Even a old outdated Chrome which is still installable on XP (cause Chrome dropped XP support some versions ago) supports modern ciphers is much better than a fully outdated IE8.
It’s not only a browser problem, you know, there are old libcurl’s and JRE’s everywhere, and people assume that it’s sufficient to ‘disable certificate check’. People think they just don’t have to upgrade. Ever. And server admins consider their companies as one prisoner in the famous dilemma, with the customer as the second one.
Users who don’t update from XP split into various groups. Some just don’t want to spend the money, others don’t want to lose support for their old table scanner or printer, and there is of course a lot of business still running on XP which has never been planned to be upgraded, like money drawing machines. Others just. don’t. care. It’s freaking working, so why should I change at all, they think. It’d be up to the content providers, the admins, to decide they should get rid of bad habits. Only if a user sees he’s on a really old boat, he will actually switch.
Perhaps web sites should finally start to indicate red banners on top of their content like “You’re browsing with an encryption method this server will stop supporting on 25th of October, 2016. Please find yourself a new browser or upgrade your OS, when you are at it“.
just for info
https://sweet32.info/ wrote XP can get RSA AES cipher with MS10-049, which is not correct ( "the cake is a lie" rofl, if someone knows this quote)
The Windows XP operating system with security update MS10-049 supports AES-based ciphersuites
But first of all this is not installed automatically and link in MS article is dead. You will find this patch in MS catalog, but it have no effect to ciphers.
So a fully patched XP Pro SP3 = no RSA AES+ MS10-049 (KB980436) manually + reboot = no RSA AES
But if you (which is really not recommended) do the POSReady hack and download additionally 64 Patches.you will get:TLS_RSA_WITH_AES_128_CBC_SHA (0x2f)
These Ciphers are added by KB3081320 (Security Update for WEPOS and POSReady 2009 (KB3081320)) which is MS15-121.https://catalog.update.microsoft.com/v7/site/Search.aspx?q=kb3081320
The ciphers were first added to Windows XP POSReady in KB3055973. That itself is a port of KB948963 from 2008's Windows Server 2003 hotfix.
On OpenSSL web page  there is statement: But the take-away is this: triple-DES should now be considered as “bad” as RC4.
On SSLLABS.com test:
1. if using RC4 there is penalty of getting total grade B
2. RC4 cipher is marked with red INSECURE
a. Is there a plan to reduce a grade of server to B if using 3DES?
b. Is there a plan to display 3DES in Cipher Suites as INSECURE (or at least WEAK)?
What do you think? Should 3DES be graded/marked insecure the same way as RC4?
Yes, there is a plan to penalise the use of 3DES. In my view, it's fine to continue to use 3DES for Windows XP, but not for anything else.
In the above screenshot, anything RC4 based is marked INSECURE, but TLS_RSA_WITH_DES_CBC_SHA (single DES!) is "just" WEAK? Even EXPORT level encryption is considered only WEAK according to Qualys SSL Labs - Projects / User Agent Capabilities: IE 8 / XP. Shouldn't that be INSECURE as well?
Personally, I think anything that uses a 64-bit block size should be marked as Insecure alongside RC4 - it's not just weak, it's possible to recover a user's session key. Amongst the IANA-registered cipher suites I can think of, that means that RC2, RC4, DES, 3DES and IDEA should be marked Insecure, and AES, ChaCha, Camellia, SEED and ARIA should be marked Secure. I don't think it's fair to call anything that's susceptible to a practical attack "fine to continue to use" - supporting IE on XP means you should be getting an F.
Retrieving data ...