, SecurityFocus 2006-01-05
Story continued from Page 1
Yet, while the data is significantly flawed, the original story told by US-CERT's list seems to be the right one. The number of vulnerabilities reported in 2005 increased, mainly due to researchers looking into the security of Web applications. The National Vulnerability Database noted the largest increase of 96 percent from 2004 to 2005, while the Symantec Vulnerability Database saw the smallest increase of 40 percent.
While publicly reported flaws jumped, that does not necessarily mean dire prospects for home users' or businesses' security, said David Ahmad, manager for development at Symantec's Security Response team.
"Web-based vulnerabilities are all over the place and they are really easy to find--they are the low-hanging fruit," Ahmad said." We have had high-profile vulnerabilities, but that is not what is driving this increase."
Finding those flaws does not require much skills, said Brian Martin, content manager for the OSVDB.
"We are seeing people discover vulnerabilities in software with tiny distribution and low installed base--free guestbooks that are written left and right, available by the thousands," he said. "And we are seeing that it takes no skill to find vulnerabilities in these applications."
Disparate dataThe number of vulnerabilities entered into four major databases vary widely over the past five years, but seem to indicate that 2005 was a banner year for bugs.
Yet, the entire focus should not be on the rash of Web application flaws, Mell said.
The computer scientist conducted an informal survey of entries for flaws in products from well-known companies and found that six of 14 software makers had seen a doubling in the number of vulnerability reports, while another four firms saw a decrease in the number of reports. The remaining four companies reported a similar number of flaws as the year before.
"I find it amazing that large and reputable software companies are seeing a large number more flaws this year (2005) than last year," Mell said.
The database managers also cautioned that the vulnerability counts for any particular year generally do not reflect the state of secure software development, only where the research community's interests lie.
"These numbers are showing the state of practice from a few years ago, rather than what the current state of practice is today," said Jeff Havrilla, team leader of vulnerability analysis at the CERT Coordination Center.
Making the issue more difficult, several software vendors move to release patches on a specific day has resulted in most security bulletins detailing multiple vulnerabilities, a situation that makes the true number of flaws harder to count, Havrilla said.