Microsoft gets information about vulnerabilities from the same sources as everyone else. They outspend every other software vendor by something like 4:1 on outside software security consultants. If they are in a privileged position regarding WinAPI software vulnerabilities at all, it is a marginally privileged position. No security person working at Microsoft would tell you they were confident that outsiders weren't holding severe, exploitable vulnerabilities back.
NSA, meanwhile, is as competent at sourcing vulnerabilities as any organization on the planet. They have internal research teams that generate them that are presumably competitive with any private research team, and they apparently purchase vulnerabilities like everyone else --- not from Microsoft, but from research teams that sell vulnerabilities.
Microsoft gives pre-release information about vulnerabilities to lots of different organizations; for instance, the IDS and network security vendors get pre-release info to create signatures. This program is, IIRC, over a decade old.
NSA is a dual-role organization; it also houses the USG's center for defensive technology expertise. It is the opposite of surprising that NSA would have the same relationship with Microsoft as, say, Symantec would.
Finally, CISPA does nothing resembling what you claimed it does. CISPA is opt-in; it cannot be used to force a company to disclose anything. CISPA is about incident data, not vulnerabilities. It is already lawful to share vulnerability information with the government. The gray area in data sharing is non- anonymized incident data, which can be covered by any of 10+ different regulations that make even IP-level metadata risky to share for collaborative defense.
CISPA is an extraordinarily short bill; you can simply read it instead of taking my word for it.
This is a publicly disclosed, publicly available program. Implying that it's somehow a government conspiracy is lying.
The original email text is unavailable as far as I can see. It of course makes perfect sense that, at least under certain circumstances, and this was the sense of limit inferable from the email, that government agencies should be given the opportunity to assess whether the item being notified about has some security implication.
The claim is made by Bloomberg, by reference to "two unnamed government officials", that Microsoft is aware that such information might be applied for reasons not primarily connected to domestic defensive security. But this is only an unsubstantiated assertion.
The number of potential exploits that are known only to Microsoft at the time of notification to those agencies would be, at a lazy guess, somewhat proportional to their exploit assessment man hours, compared to the overall exploit discovery effort. I would think that would be the much smaller proportion.
1. http://mobile.bloomberg.com/news/2013-06-14/u-s-agencies-sai...
"Microsoft helped the NSA to circumvent its encryption to address concerns that the agency would be unable to intercept web chats on the new Outlook.com portal;" [1]
[0] https://twitter.com/ggreenwald/status/355391355939340288
[1] http://www.theguardian.com/world/2013/jul/11/microsoft-nsa-c...
Your message is a red herring; I could debate your conclusions, but what would be the point? I'm sure you've got 10 more red herrings up your sleeve.
To be fair, there is not much daylight between: 'The root comment suggests that Microsoft arranges to provide NSA with vulnerabilities that they deliberately do not patch' and what the Guardian reported about Outlook. At least to me, but I'm not a security expert. Is it not considered a 'vulnerability' if Microsoft hands over the keys so to speak?
Also on a side note, it is rude to call me out like you did. I don't have herrings up my sleeves. I don't care about internet points. I just wanted to contribute more information to the discussion.
I don't doubt it happens, so much as I expect extraordinary claims be backed with extraordinary evidence.
Very frustrating as I work with a Microsoft-oriented company at the moment. Any mention of this to their architectural team results in nothing short of "mwuhahaha you're talking shit". There is some weird universal trust there that really makes no sense at all.
To add insult to injury they don't log, don't have an IDS and don't have a clue stick to hit themselves with.
Their funeral!
You can spin it as a secret NSA-program, but I don't think that's very useful as this program does legitimately help a ton of users.
(Not that it really would excuse things, but I'd be very surprised if intelligence agencies don't have people inside companies like Microsoft anyways.)
I've worked for a few large European financial orgs over the years and MS will not disclose issues early to them despite begging and piles of money. We're talking 50k+ employees here.
However DoD connected companies like Lockheed Martin, Boeing and Raytheon are right in there without an invite.
You just get the feeling there's something off when relationships are made like that.
Also, mwuhahaha you're talking shit.
Not impossible for NSA to get in but a lot more difficult.
None, unless you install them and activate them? What sort of desktop distro comes with pre-configured webservers?
It's a lot easier to attack the website running on the server and escalate from there then attack the OS directly.
So the vulnerability would have to be in something that both servers and desktops have in common.
You're cute.
And there's enough people with such a key to choose from: https://www.archlinux.org/master-keys/.
(I run Arch as well, but I have no such illusions of security.)
Edit: not to discredit Linux of FLOSS, at least there is the possibility to analyse the source code.
Where there's a will there's a way especially if you have billions of dollars in funding and the freedom to do so i just aim to make it as difficult as possible.
There's a lot of things you can do to protect you privacy but it all starts with a good choice of OS and hardware.
If the foundation is compromised there's no point in anything else you do to protect your privacy.
Edit: Yes Arch it's not particularly security conscious I choose to compromise some security to stay on the bleeding edge now that could mean that I get some bad code sometimes but that also means that it gets fixed sooner too.
Having a rolling release system that you can mold to your needs is worth it for me.
So it runs on i686 and amd64 only. How do you select "good" hardware?
However since there's no such thing you will have to choose who you trust.
http://www.bloomberg.com/news/2013-06-14/u-s-agencies-said-t...
Every single one of these vulnerabilities could be seen as a backdoor, except Microsoft can have plausible deniability, since they are not actually putting a backdoor in the OS themselves - they're "just telling NSA about the vulnerabilities that exist".
If something like CISPA passes, which NSA keeps pushing for, this capability will expand dramatically, as all companies will be forced to give these vulnerabilities to NSA, but not to "protect us" and for cyber "security", as they keep claiming when they try to promote laws like these, but for offense. They will hoard every single one of them, and then use them in such automated systems to infect millions of computers.