Cool Solutions

Vista is Most Secure OS: I Can’t Stop Laughing…


January 30, 2008 2:40 pm





Setting the Stage

Microsoft has released a self-analysis of Vista that proclaims among other things that Vista has less found vulnerabilities than any other OS in their first year. Written by a member of their Trustworthy Computing group, this report exhibits more spin than the teacup ride at Disney Land.

It’s an interesting way to get a favorable number — to say that the code base is brand new, thereby resetting all the expectations. Wrong way, guys, newer code always has more bugs in it, you just counteracted the whole point. Never mind the fact that Vista’s feature set could arguably (and kindly) be called XP Service Pack 4, and totally ignoring the fact that it took 5 years to come up with and release this new set of code/features/functionality

We’re Not as Bad and Getting Better

Even more amusingly is the continuing notion that a lesser number of found flaws somehow equates to a more secure code base. This ignores some basic software concepts, such as the fact that every code base has flaws, it’s just how many flaws there are and when/if they are found that’s the difference.

The “Security through Obscurity”, or “Our code sucks less” mentality isn’t new, but it’s about as accurate a measurement as body counts in Vietnam were, and about as good an indicator of winning the war on bugs and flaws.

In order to make a good comparison, let’s see what the security research community has to say about numbers of bugs in lines of code:

Commercial software typically has 20 to 30 bugs for every 1,000 lines of code, according to Carnegie Mellon University’s CyLab Sustainable Computing Consortium.

Doing a Number(s) on Windows

There exists an excellent whitepaper and discussion of Security Vulnerabilities in Software Systems from Colorado State by Alhazmi, Malaiya and Ray that clearly lays out a formulaic method for determining bugs in code, which is very helpful in understanding the issues of bugs in code. WordPress isn’t the easiest to do mathmetical notations in, so I’ll spare you the nifty formulas and just use a calculator and some common terms to explain the deal.

Note: LOC = Lines of Code, SLOC = Source LOC, KLOC = 1,000 LOC and MLOC = Million LOC and BKLOC = My own notation for Bugs per KLOC.

According to published sources, Microsoft XP contains 40,000,000 lines of code, so let’s do the math to see how many possible bugs in XP total:

40,000,000 / 1,000 (chunks of code) = 40,000
x 25 (bugs per chunk) = 1,000,000

Even taking the number of possible bugs down to the lowest approximation (20) nets you almost 800,000 possible bugs in Windows XP.

Since the report is about Vista, let’s do the same analysis for the 50,000,000 lines of code in Vista to see how many possible bugs total:

50,000,000 / 1,000 (chunks of code) = 50,000
x 25 (bugs per chunk) = 1,250,000

Let’s take Microsoft’s word and assume (very generously) that they’ve improved their security dramatically, so for the sake of argument we can cut the number of possible bugs in half, getting:

50,000,000 / 1,000 (chunks of code) = 50,000
x 12.5 = 625,000

Jeffrey Jones’ report shows us that there have been 66 flaws reported in the first 6 months of Vista’s life, so extrapolating that out to 112 in a year, it’s going to take a while to get significant buggage out of Vista.

Doing a Number(s) on Linux/OSS
The aforementioned Stanford University study comparing security between proprietary and open source code found:

According to a four-year analysis of the 5.7 million lines of Linux source code conducted by five Stanford University computer science researchers, the Linux kernel programming code is better and more secure than the programming code of most proprietary software.

The number of possible vulnerabilities per 1,000 lines of code for the Linux Kernel?

The study identified 0.17 bugs per 1,000 lines of code in the Linux kernel.

The latest Kernel 2.6 project numbers from the project pages puts the number of lines of Kernel Source code at about 10,000,000, so let’s do the numbers:

10,000,000 / 1,000 (chunks of code) = 10,000
x .17 (bugs per chunk) = 1,700

Ah, but you may say we aren’t comparing the same number of lines of code, and that’s true, so let’s do just that:

50,000,000 / 1,000 (chunks of code) = 50,000
x .17 (bugs per chunk) = 8,500

There you have it, even if the Linux Kernel were 5x it’s size, there would still be less than 10,000 bugs.

Other Comparisons
It’s not just the Linux Kernel that’s been analyzed for security, the LAMP (Linux, Apache, MySQL, PHP) stack has undergone scrutiny for bugs per lines of code, and Coverity recently did a source code analysis of 32 Open Source projects with an average bug density of .43 bugs per thousand lines of code, with the LAMP stack scoring better than average when compared to the other projects.

Let’s not forget the Great-Grand-Daddy of all code comparisons, “More than a Gigabuck: Estimating GNU/Linux’s Size” by David Wheeler. This paper and it’s conclusions did a lot to help convince and demonstrate the value of the Open Source Community and development process.


Let’s call it what it is: Vista, even if it’s only 1/2 as buggy as XP is still vastly more buggy than the comparable number of lines of Linux Kernel or Open Source code. Microsoft has admittedly come a long way with their security improvements, but the numbers tell us that there is quite a distance to go before Vista can be proclaimed as being more secure or less buggy than Open Source code.


0 votes, average: 0.00 out of 50 votes, average: 0.00 out of 50 votes, average: 0.00 out of 50 votes, average: 0.00 out of 50 votes, average: 0.00 out of 5 (0 votes, average: 0.00 out of 5)
You need to be a registered member to rate this post.

Categories: Uncategorized


Disclaimer: This content is not supported by Micro Focus. It was contributed by a community member and is published "as is." It seems to have worked for at least one person, and might work for you. But please be sure to test it thoroughly before using it in a production environment.