, SecurityFocus 2007-01-17
Flaws in Web applications boosted the bug counts for 2006 by more than a third over the previous year, according to data obtained by SecurityFocus from the four major vulnerability databases.
On Monday, the Computer Emergency Response Team (CERT) Coordination Center released its final tally of the number of flaws the organization processed in 2006. Counting both public sources and private submissions directly to the CERT Coordination Center, the group logged 8,064 vulnerabilities last year, an increase of 35 percent over the number of flaws reported in 2005.
The three other major flaw databases--the National Vulnerability Database, the Open-Source Vulnerability Database, and the Symantec Vulnerability Database--recorded jumps anywhere from 20 to 35 percent in 2006 compared to 2005.
The greatest factor in the skyrocketing number of vulnerabilities is that certain types of flaws in community and commercial Web applications have become much easier to find, said Art Manion, vulnerability team lead for the CERT Coordination Center.
"The best we can figure, most of the growth is due to fairly easy-to-discover vulnerabilities in Web applications," Manion said. "They are easy to find, easy to create, and easy to deploy."
The burgeoning flaw counts for 2006 should come as no surprise to most security researchers. A jump in the number of vulnerabilities recorded by the same four databases in 2005 had also been blamed on easy-to-find bugs in Web applications. In the first half of 2006, more than three quarters of all software flaws affected online applications, according to security firm Symantec, the owner of SecurityFocus. And a report released in October by the Common Vulnerabilities and Exposures (CVE) Project found that the top-three categories of flaws were specific to Web programs and accounted for 45 percent of the bugs reported in the first nine months of the year.
Simply searching through source code or using Google code search can turn up a large number of potential security issues, allowing even novice flaw finders to pinpoint possible security holes. The maintainers of the flaw databases have been inundated with submissions found by would-be vulnerability researchers who use simple string-matching programs find potential issues in open-source applications, said Steven Christey, the editor of the CVE Project maintained by The MITRE Corp., a non-profit government contractor.
"Many people are doing 'grep and gripe' research," Christey said, referring to the flexible search program grep commonly part of Unix-like systems. "They are doing a regular expression search, looking for patterns. If they get a match they will report it to the public, but sometimes what ends up happening is they are reporting false positives."
While novices are focusing on Web applications, other researchers have started focusing on other parts of the operating system as well as popular applications. Tools, known as fuzzers, have become an increasingly popular way to check software for problems caused by the input data given to the program. Such tools have been so successful in finding flaws that some researchers have resorted to the controversial tactic of releasing a bug every day for an entire month to garner attention to the issues.
"You have an emerging levels of sophistication for vulnerability researchers," Christey said. "You have a lot of people who are able to find the low-hanging fruit. But for major software, it seems to be getting more difficult for top researchers to find these issues--they have to work harder, spend more time, spend more resources, (and) do more complex research."
Web flaws boost bug countsEasy-to-find vulnerabilities in Web applications significantly boosted the number of flaws found in 2006, exceeding the previous year by anywhere from 20 percent to 50 percent.
Sources: Computer Emergency Response Team Coordination Center (CERT/CC), National Vulnerability Database, Open-Source Vulnerability Database, and the Symantec Vulnerability Database.
The surge in vulnerabilities in 2006 does not necessarily mean that the Internet is a less safe place for computer users.
Many of the Web applications in which flaws were found are community projects not typically used by major companies, said Brian Martin, content manager for the Open-Source Vulnerability Database.
"For the personal sites and the mom-and-pop stores that rely on the software, it certainly affects them," Martin said. "But larger companies likely aren't affected."
Applications written in the popular dynamic Web programming language, PHP, appeared to account for 43 percent of the total vulnerabilities reported in 2006. The language is typically used in community-created software and smaller Web sites, but a number of notable Internet giants, such as Yahoo! and Google, also use PHP.
While Web applications may account for the boost in vulnerability numbers, the smaller number of flaws found in operating systems and client-side applications typically pack a bigger punch.
"From a core operating system standpoint, we are more secure, but the reality is that malicious code has not gone away," said Oliver Friedrichs, director of security response for Symantec. "Malware is still getting on your system, it is just not using core operating system vulnerabilities to do it."
And, while they make up a small fraction of the overall number of vulnerabilities, previously unknown--or zero-day--flaws targeted by active attacks became a big trend in 2006, Friedrichs pointed out.
"The real threat with a zero-day is they are frequently used in a very targeted fashion against companies and enterprises to steal information, where a simple web vulnerability on the Internet does not have as much of a material impact," he said.