Slashdot Mirror
The Annual US-CERT FUD Festival
Joe Barr writes "Joe Brockmeier and I have teamed up in a story on NewsForge to point out how the mainstream and trade press misrepresent the annual summary of vulnerabilities from US-CERT. They're doing it again this year to make it appear as if it is more secure than UNIX/Linux. Pamela Jones did a similar report at Groklaw over the weekend." From the article: "One figure represents the vulnerabilities found in Windows operating systems: XP, NT, 98, and so on. The other represents a total figure not just for Solaris, AIX, HP-UX, the BSDs, and Linux, but for a hundred different versions of Linux. The sum of all the unique vulnerabilities from all the Linux distros does not equate to the sum of vulnerabilities in any single Linux distro, and one could say the same about the various versions of Windows. That's why it is a completely meaningless exercise to discuss those totals as if they present an accurate picture of the relative security of Windows and Linux. " We've reported on the US-CERT list already this year. NewsForge is a sister site to Slashdot.org, both of whom are owned by OSTG.
It's equally unfair to lump Windows 98, NT, 2000, XP all together. They could be looked at as different "distros" of Windows. Should pick the best or latest OS from each group with the least vulnerabilities to compare.
Shouldn't we be asking the more pertinent question: why do all the various operating systems have so many vulnerabilities? When it comes to such things, this shouldn't be a competition. OS builders should be striving for zero tolerance to vulnerabilities and there shouldn't be an quibbling over the number that exist.
GetOuttaMySpace - The Anti-Social Network
But that's not the same - we're talking about basically one Windows product with its associated unique vulnerabilities, but when we talk about Linux distros, we talk about several different ones that have the *same* vulnerability counted multiple times because it exists in multiple distros. Just one look at the CERT list and you will see all the duplicates in there. And then of course, even if you remove the duplicates, you are still left with vulnerabilities that were only present in one distribution, but got counted against "Linux/Unix" although 99% of the distros were never affected.
Part of the contention is the repeat entries with the "updated" notation. So if you throw out all 141 "updated" occurrences in the Microsoft section, that leaves 671 (812-141=671).
If you throw out all 1437 "updated" occurences in the linux/unix secion, that leaves 891 (2328-1437=891). Subtracting Apple OS X (130) and Sun Solaris (77), Linux/Unix ends up with 13 more vulnerabilities than Windows (891-130-77=684), but it's for more operating systems, so it may be fair to divide that 684 further.
The theory of relativity doesn't work right in Arkansas.
Good point and I'd like to add, What about the time length between when vulnerabilities are found, and then patched? Surely, they thought about that. Linux and Unix can continue to have more "reported" vulnerabilities than Windows, but if they are patched faster than Windows, doesn't that count for something?
I have nothing clever to put here...
They're doing it again this year to make it appear as if it is more secure than UNIX/Linux.
What is "it"? Slight tinge of paranoia here, maybe?
Let's review the score here:
- It does not matter what material is published, the fact of the matter is that every Windows PC in the world regularly has visible and non-trivial security issues, while on Linux and OS/X these issues are generally theoretical.
- People's perceptions of Windows are very simple: it's a piece of crap that they use because it came with the box and everyone else uses it.
- The relative security of Windows vs. the World is not a deciding factor in most people's use of Windows. It's largely a captive, neutered market.
- For people who actually do care, no amount of statistics can change the visible and perceived situation. When I choose to ban Windows in my company, it's not because I read some website or article. It's because I'm sick and tired of removing spyware from people's PCs.
Complaining about these statistics is to give them credibility. Those who chose on the basis of security will ignore this data, and those who chose on other criteria won't care about this data.
My blog
I'm an automation officer in the U.S. Army, and I know for a fact that we're full of Microsoft shills and contractors with Microsoft loyalties. We don't employ Unix/Linux in an enterprise manner; the government sold its soul to Microsoft years ago. Unix is used on some Army tactical platforms, though. Food for thought.
I would have expected better *if* CERT was still in the hands of a university. I wouldn't trust a government analysis as far as I could throw a CRAY.
Considering Linux is a Kernel, to say there were 1000s of bugs again Linux is silly.
It would be interesting to see all of the Windows application vendors lumped into the "Microsoft security flaws" category in a similar manner. I've seen quite a few Windows applications from all sorts of software vendors with issues this last year and noticed they weren't listed. While one might argue at first that this would be unfair because of all of the commercial products available for Windows, I'm not sure Windows wouldn't still have an advantage. Just go to sourceforge.net and start counting up all the projects available there that could be lumped into Linux "security flaws."
Looking just at core operating system applications, Fetchmail doesn't make the cut. In fact, it's inappropriate to include GCC in there since I'm certain they didn't include Microsoft development environment tools in the Microsoft count. An apples-to-apples comparison isn't appropriate and perhaps for those uneducated technical journalists that like to make comparison stories, a kernel-to-kernel, browser-to-browser (e.g. IE vs. Mozilla vs. Opera), office suite to office suite, and other category-based comparison is the only appropriate approach.
Whats worse is the fact that a POP3 Client Buffer Overflow on Windows would not be included at all as one doesn't ship with Windows. Linux distros generally ship with thousands of clients and servers while Windows ships with the bare minimum. To do a true security comparion you would have to compare either just kernel exploits with OS exploits, then compare all popular software for windows with all popular software for Linux side by side in a catagory basis (POP3 clients being a catagory)
I vote for the "solves-my-problem-but-not-yours" distribution, which is clearly the best.
Incidentally, I am also in favor of settling on ONE (1) tool for all mechanical uses.
I favor the two-handed hewing axe, but I might be persuaded to vote for the claw hammer.