NEWS  
< http://www.securityfocus.com/news/6568 >  
 
--------------------------------------------------------------------------------
 
   
Panel Probes the Half-life of Bugs
 
By Kevin Poulsen, SecurityFocus Jul 30 2003 7:41PM  
 
LAS VEGAS--Software security holes never die, they fade from the Internet at a rate of 50% every thirty days after a patch is released, according to the results of a study released at the Black Hat Briefings security conference here Wednesday.  
 
Researchers at the California-based security company Qualys analyzed the results of 1.5 million vulnerability scans conducted since January 2002 by the company's free and commercial vulnerability scanning services. The results put some statistics behind what security pros and hackers have known anecdotally for years: new vulnerability announcements tend to kick off a kind of festival of hacking, in which anyone who gets their hands on an exploit can have their way with innumerable systems around the world, as companies and individuals slowly adopt the fix.  
 
Currently in that hacker sweet spot is the Microsoft Windows DCOM RPC hole announced July 16th. Using exploits that have already been released to public mailing lists, and others being swapped privately in IRC, anyone with the inclination can tap into computers running most versions of Windows, if the target system hasn't been patched and isn't behind a firewall.  
 
According to Qualys' analysis, the RPC vulnerability is now the most prevalent on the Internet.  
 
By an odd quirk, some bugs violate the 30-day half-life rule and actually start growing in prevalence about 90 days into their lifecycle, said Qualys CTO Gerhard Echelbeck, presenting the research in a panel discussion. "There are a number of vulnerabilities that come back." Echelbeck attributes this bug-immortality to network administrators cloning new systems from old, unpatched images.  
 
The researchers also found that 80% of exploits are released within 60 days of a vulnerability's announcement.  
 
"This work on the statistics validates what we've always considered to be a rule of thumb," said Black Hat organizer Jeff Moss.  
 
Disclosure Debate  
Security researcher Mark Loveless, also on the panel, said the computer underground knows well how to take advantage of the opportunities presented by the patching curve. "There's a lot of awareness already in the underground culture, and they have already been ramping up their efforts to a large extent, because they know they have a very short window in which to work," said Loveless. "There are people who work extremely fast at this"  
 
Software vulnerabilities are emerging as the unofficial theme of this year's Black Hat Briefings. Wednesday morning the Organization for Internet Safety held a panel discussion on its official guidelines for vulnerability reporting and response, released Monday. The guidelines would give vendors at least 30 days to produce a patch for a vulnerability before a bug-finder goes public with it. The bug-finders would then withhold exploit code and technical details for another 30 days after the advisory.  
 
Chris Wysopal, research director for @stake and an author of the guidelines, said full and immediate public disclosure of security holes was once the only way to get vendors to plug them, but that today the practice does more harm than good.  
 
"The environment has changed in the last seven to ten years, when the Full Disclosure movement started," said Wysopal. "The vendors at that point weren't really communicating with anybody who wasn't a big customer, and didn't have any process at all for dealing with security issues.  
 
"Things don't work that way now," said Wysopal. "At some point vendors started to realize that they had to take this stuff seriously"  
 
Some audience members were skeptical, and charged that the guidelines were aimed at increasing the value of early-notification services that some security companies provide, in which paying customers get early access to vulnerability data under a confidentiality agreement. The guidelines neither block nor encourage the practice.  
 
Others worried that vendors would use the guidelines to delay fixing security holes. Scott Culp, security manager at Microsoft, an OIS founding member, countered that the guidelines would actually put software-makers' feet to the fire, because it mandates that vendors give bug-finders regular updates on their progress in patching a bug. "From a vendor perspective, these guidelines don't let us off the hook," said Culp. "Far from it."  
 
Much of the talk at the conference has focused on the Windows DCOM RPC hole, which many expect to be the target of an Internet worm in the days or weeks to come.  
 
The hole affects most versions of Windows, including Windows Server 2003, which ostensibly had been the subject of an intensive pre-release code review as part of Microsoft's 18-month-old "Trustworthy Computing" security initiative.  
 
In congressional testimony last week, Microsoft chief security strategist Scott Charney told lawmakers that the company had spent $200 million trying to make Windows Server 2003 secure. In a prepared statement, he called the new bug "disappointing," but offered little hope that it wouldn't happen again. "Such occurrences are part of major operating system development," said Charney. "These systems -- in all platforms, including Windows, Linux, and Unix -- will always suffer vulnerabilities."  
 
 
 
   
 
<tips@securityfocus.com>
 
 
 
--------------------------------------------------------------------------------
 
Copyright © 1999-2003 SecurityFocus

iptables -P FORWARD -j ACCEPT