Users Wonder If Microsoft Exec's Security Claims Are Trustworthy
Jaikumar Vijayan Today’s Top Stories or Other Security Stories
September 11, 2006 (Computerworld) -- BOSTON -- Since Microsoft Corp. launched its Trustworthy Computing initiative in January 2002, the company has substantially improved the security of its products, its security chief said here last week.
However, some Microsoft users at The Security Standard conference were less charitable in their assessment of the company's progress. Most of those attendees concluded that it's too soon to deliver a verdict on Microsoft's security efforts.
During a panel discussion, Ben Fathi, corporate vice president of Microsoft's security technology unit, gave the security capabilities of today's Microsoft products a grade of B+. Five years ago, he said, that grade would have been a C- or a D.
The improvements are the result of a "cultural shift" at Microsoft sparked by the Trustworthy Computing memorandum that Chairman Bill Gates issued nearly five years ago, Fathi said.
Since then, the company has completely overhauled its product development processes, and it has trained developers to write secure code, Fathi explained. The initiative prompted the creation of the Software Development Lifecycle (SDL), a security-focused quality-control process that's now used for every software product developed at Microsoft.
"Last year, we had over 300 products that went through this, and with three exceptions, they all passed," Fathi said. The products that failed were blocked from release until the development team fixed the problems or found a way to mitigate them, he said.
"Vista is the first product that has gone through SDL from inception to end," Fathi noted. As a result, he said, "there is a whole lot of improvement to security in the product" compared with earlier products.
Users at the conference were skeptical of that claim, noting that it will take time for Microsoft to demonstrate that products such as Vista are indeed more secure.
Mark Olson, manager of information security at Beth Israel Deaconess Medical Center in Boston, said that so far at least, Microsoft appears to be on the right track with its Trustworthy Computing initiative. However, he also noted that the health care organization has yet to deploy any products built using Microsoft's SDL process.
"Until [such a product] is deployed and we have a solid 18 months of runtime, I won't know for sure if they have gotten any better," Olson said, adding, "It's wise to be wary of sales pitches."
Still, he noted, Microsoft's patch release process and its efforts to develop tools such as its rootkit detector have been positive steps.
Holding Off
Marcin Czabanski, chief security officer at Medical Network One, a Rochester, Mich.-based provider of managed health care services, said he is taking a "wait and see" approach to determining whether the security of Microsoft products has truly improved, as the company claims.
"There have been a lot of improvements of their processes, and their products are more stable," Czabanski said.
Even so, Medical Network One plans to hold off on deploying Microsoft's upcoming Windows Vista operating system across the enterprise until there is reasonable evidence that it's secure.
Czabanski also suggested that Microsoft needs to improve the reliability of its patches. He noted that there have been occasions in the past where Medical Network One has had to reinstall Microsoft patches that were initially faulty.
"They have been going in the right direction, but there's room for improvement," added Randy Bachman, information security principal at Lockheed Martin Information Technology in Newington, Va.
For instance, Bachman said, although Vista appears to be more secure than its predecessors and has integrated useful functions such as data encryption, Microsoft should also be working to add improved integrated spyware-detection and antivirus functions.
AI
