SecurityFocus published an interview with BlackHat/Defcon speaker Dan Kaminsky. He was guest-hacker at the Microsoft BlueHat event, and met Redmond’s engineers and management. His conclusion: “My sense is that a combination of respect for SP2 and growing fear of Google (which has an entirely different, and arguably more managable security posture than Microsoft can achieve) has really pushed people towards seeing security in 2005 as stability was in 2000/2001.”
Dan Kaminsky on Microsoft’s Security
30 Comments
2005-07-21 1:38 pm
Thom Holwerda
Pravda– it’s probably not coincidental that your name is also the name of the Communist newspaper that was known for spinning the truth to better fit the communist views 
.
It is possible to achieve security with opennes. But that does not mean that a system can only be secure when it’s open. I’ll repeat the Joe story. Read it carefully: “Joe puts up an umbrella when it rains. Joe puts up an umbrella. Does that automatically mean that it’s raining?”
Logic, sonny. It’s a difficult concept, but in the end you’ll get it.
2005-07-21 1:48 pm
2005-07-21 2:17 pm
Budd
Pravda– it’s probably not coincidental that your name is also the name of the Communist newspaper that was known for spinning the truth to better fit the communist views .
Thom,you are pathetic.Even if you intention was to joke. It sucks.You brought no argument into discussion except pointing that that users name is pravda. Get a life.
2005-07-21 1:30 pm
2005-07-21 1:53 pm
2005-07-21 2:06 pm
-
2005-07-21 2:09 pm
2005-07-21 2:15 pm
2005-07-21 3:11 pm
2005-07-21 3:21 pm
2005-07-21 4:12 pm
Anonymous
Pravda you know something…you are a joker. How old are you? All you are doing here is insulting yourself and your parents by acting so childish and foolish. If openness == security then no exploit on Linux will exist. How about firefox bugs? What about big history of sendmail attacks? Don’t forget buffer overflow was mostly used on UNIX initially.
Now please stop your childish rant pravada and learn to grow up. You will never get to see the source code of Windows because hey it is something called IP…intellectual property and if you are so intelligent then go write your own OS
2005-07-21 4:29 pm
2005-07-21 11:01 pm
2005-07-22 7:44 am
2005-07-22 8:40 am
And we got rid of Windows stability problems after 2001, right?
> And we got rid of Windows stability problems after
> 2001, right?
Yes. After 2000, actually, or to be more precise, after the release of Windows 2000.
> And we got rid of Windows stability problems after
> 2001, right?
Yes. After 2000, actually, or to be more precise, after the release of Windows 2000.
Umm… Windows 2K, in its RTM form was far from being “rock solid”. Same goes for Windows XP.
At least on my SMP setup(s), Windows 2K became stable (read: NT4SP5 like) right around SP3. XP/SP2 has yet to achieve this goal.
Windows 2000 wasn’t used by home users, though. If we’re including Windows 2000, we might was well include NT workstation and say the stability problem was solved in the mid 1990s. To be more accurate, the stability problem was solved generally in 2002, with the release of WinXP.
i cant realy comment for others then myself but i have been running win2k as a desktop for years now…
I don’t know what Microsoft even thinks. Until Microsoft open sources their entire OS, entire toolchain, and all the CPU and other chips, there will be no security on Windows.
The reason Microsoft bought their own anti-virus and anti-spyware is so Microsoft’s rootkits can still work undetected (along with Microsoft’s spyware applications) and Microsoft can give security/spyware/adware exceptions to those it has to and those who pay for an exception.
The computer, as defined by Microsoft and Intel, has proven to the public that it cannot be trusted. No matter how much big brother TPM hardware Microsoft/Intel adds, it will not help change this belief.
And so it just makes me laugh.
Trusting Microsoft with security is like trusting the fox to guard the henhouse.
the problem is the default, out of the box, permissions setup from ms. until an ordinary user is not merrily chugging along as an admin, things can only be covered up with band aids.
one can argue that you can run as a limited user, etc., but that only happens for those in the know. tons of biz and most home users have no idea and use ms default of admin (with no password to top it off).
the priviledge model is due to change with longhorn, and when it does, ms achilles heel will be gone.
and yeah, i agree windows has been rock solid since the release of windows 2000. 2000, xp, and server 2003 perform fantastically. the lockups and blue screens of the windows 95/98 era are gone.
I was yet to see any problem or even bluescreen when people worked as users.
Personally I didnt have a bluescreen for years, and spyware only once in the last 5 years (and that was ironically when I logged in as admin just to download and install an update on the fly…).
I got my linuces and my mac to core dump or lockup pretty quickly though. (yeah I know, this is against ‘the common opinion’ – only windows gets fucked up over time, unices are rock-solid… well, not for me)
The thing is, the Windows NT line is a very capable and decent operating system, with a lot of security principles that are very much the same as on UNIX. But as Anonymous (IP: 68.235.182.—) pointed out, the default setup leaves a lot to be desired.
A couple comments:
1. Better default setups, while most definitely needed, are a not a substitute for proper user education. If the user isn’t properly informed, they’re just as likely to turn off the new security measures as they are to use them.
2. The 3rd party software vendors are at least partly to blame for the tendency of Windows users to run as administrators. A lot of software packages are so shoddily coded that they refuse to run under a non-administrative account.
A brief restatement of the message that Microsoft (or their minions) voted down:
Trusting Microsoft with security is like trusting the fox to guard the henhouse.
It seems any sort of non-positive statement about Microsoft is getting voted down. And Microsoft complains about bias. Reminds me of an old Polish proverb.
What voting really needs is a list of who voted along with their IP addresses and a smart lookup of where those IPs are coming from.
No, your childish ignorance enriched with mindless zealotry was voted down:
Until Microsoft open sources their entire OS, entire toolchain, and all the CPU and other chips, there will be no security on Windows
A completely open chain is a well accepted precept of secure systems. If you are ignorant of this, that is your own blindness.
And it doesn’t change the fact that we need voting records.
The world knows very well that Microsoft buys decisions from Congress, buys analyst studies, etc. I would hardly be surprised if Microsoft and their lackeys are perverting the voting system here.
In a somewhat ironic fashion, the lack of a voting record underscores the very issue we are talking about. Without a completely independent and 100% thorough verification of what is secure, there is no security. The same with voting.
I voted your comment to -5 with one click, and as far as I can recall, I don’t have a job contract with Microsoft.
A completely open chain is a well accepted precept of secure systems. If you are ignorant of this, that is your own blindness.
Great. But being open doesn’t guarantee a secure system. False logic, Pravda. Kindergarten stuff: Joe puts up an umbrella when it rains. Joe puts up an umbrella. Does that automatically mean that it’s raining?
I will relate the following:
In 1984 the co-author of Unix Ken Thompson gave a theoretical example of how even an open toolchain could be compromised He described what he called the “cutetst program I ever wrote” in a paper subsequently known as Reflections on Trusting Trust. Thompson set about producing a program which produced as output an exact copy of its source, in other words a program that will create a self-reproducing program. His compromised compiler could be used to replace the standard UNIX compiler, and the intruder could then remove the traces from the compiler source code, knowing that any program generated by the compiler would harbor the Trojan.
Thompson’s point was simple. “You can’t trust code that you did not totally create yourself. No amount of source-level verification or scrutiny will protect you from using untrusted code,” he pointed out.
The answer is to have a clean room environment – both tool chain and operating system – that the user has created and knows is free of compromises. This can then be trusted to bootstrap the toolchain and OS of your choice.