Study: Largest vendors account for fewer software flaws
Improved quality assurance, testing likely reasons for improvement
July 25, 2007 12:00 PM ETComputerworld -
Though it might not seem that way, the top 10 most vulnerable software vendors -- and, yes, that includes Microsoft Corp. -- are contributing a smaller percentage of all vulnerability disclosures per year compared with five years ago.
That's according to an analysis by Gunter Ollmann, director of security strategies at Internet Security Systems Inc.'s X-Force team, which is a unit of IBM.
Ollmann, who crunched vulnerability data gathered by X-Force between 2002 and 2006, said the overall percentage of security flaws disclosed by the most vulnerable software vendors dropped from 20.2% in 2002 to 14.6% during that period.
Much of that decrease is likely the result of improved quality assurance and testing processes by the most vulnerable software vendors, Ollmann said. Most of their software packages have been through multiple versions and have been combed thoroughly for vulnerabilities by security researchers, he said.
According to Ollmann, the most vulnerable vendors typically have also been the biggest software vendors and those with the largest installed bases. Traditionally, security researchers and hackers have gone after vendors with the biggest installed bases because that is where they can have the biggest impact.
As larger vendors begin to do a better job of locking down their software, hackers and software researchers have begun focusing their attention on newer vendors and their applications, which has resulted in an overall increase in the number of vulnerabilities being reported, Ollmann said.
In the past five years, the list of the most vulnerable vendors has consistently included Microsoft, Cisco Systems Inc., IBM, Sun Microsystems Inc. and the Linux Kernel Organization Inc., Ollmann said. Others on the list for 2006 were Oracle Corp., Apple Inc., Mozilla Corp. and Adobe Systems Inc.
Together, these vendors accounted for 964 vulnerabilities in 2006, or roughly 14% of the total disclosed by all vendors. The remaining 85% or more were accounted for by smaller vendors. For instance, nearly 1,000 vulnerabilities were reported by relatively small vendors of PHP applications, he said.
Ollmann acknowledged that the vulnerabilities reported by the top 10 are likely to have affected a far larger number of people than a majority of the flaws reported by smaller vendors. At the same time, it is also true that the larger vendors have been doing a better job in fixing reported flaws, he said.
Out of the top 10 vendors, only 14% of the publicly reported flaws remained unpatched in 2006, while 65% of all other publicly reported flaws remain unpatched.
Larger vendors are also more likely than smaller ones to have better mechanisms for detecting flaws and alerting customers, he added.
Read more about security in Computerworld's Security Knowledge Center.
The top 10 most vulnerable software vendors
Additional Resources



Learn the important issues you must consider before starting your next mobility initiative. Get your mobility white paper from IDC now, compliments of Sybase.
White Papers & Webcasts
Death to PST Files
Download Now
The Tangled Web: Silent Threats & Invisible Enemies
Download Now
Tape Killed the IT Guy
Watch Now
Forrester Consulting Mobility Study: Taking Control of Enterprise Mobile Device Diversity
Download Now
BRM: What You Can Do To Reduce Risk In Challenging Times
Watch this webcast now!
What IT Must Do to Support Employee-Owned BlackBerry, iPhone and Android Mobile Devices
Download Now
Web 2.0, Social Media and the Dark Web - A Web Criminals Paradise?
In this discussion, learn about the challenges of protecting your users from the potentially unsafe content hidden in the "Dark Web".
eGuide: Enterprise Security
Smart Security Strategies for 2010. Read now!
Disaster Recovery 2008: Reduced Costs and Improved Performance
How long can your Enterprise afford to be without your data? With an accelerated disaster recovery program, you never have to answer this...

