News Analysis: The software maker admits to withholding details on security vulnerabilities to protect customers from bad guys, but critics say that policy increases the risk for everyone.
Microsoft has fessed up to hiding details on software vulnerabilities that are discovered internally, insisting that full disclosure of every security-related product change only serves to aid attackers.
The companys admission follows criticisms from a security researcher
that its policy of silently fixing software flaws is "misleading" and not in the spirit of Microsofts push for transparency.
In an interview with eWEEK, Mike Reavey, operations manager of the MSRC (Microsoft Security Response Center), said the companys policy is to document the existence of internally discovered flaws as well as the area of functionality where the change occurred, but that full details on the fixes are withheld for a very good reason.
"We want to make sure we dont give attackers any [additional] information that could be used against our customers. There is a balance between providing information to assess risk and giving out information that aids attackers," Reavey said.
When Microsoft receives a report of a security flaw from external researchers, Reavey said, the MSRC conducts an extensive investigation to look at all the surrounding code to make sure a comprehensive fix is pushed out the door. If a related bug is found internally, it will be fixed in the eventual patch, he said, but the details will be kept under wraps.
Read more here about Microsofts patch-creation process at the MSRC.
However, critics argue that silent fixes have a way of backfiring and hurting businesses that depend on information from the vendor to determine deployment time frames and the actual severity of the patched vulnerability.
According to eEye Digital Security, which sells host-based IPS (intrusion prevention system) technology, silent fixes from Microsoft are commonplace.
"It is the skeleton in Microsofts closet. We routinely find them," said Steve Manzuik, product manager of eEyes security research team, in Aliso Viejo, Calif.
In an interview with eWEEK, Manzuik said Microsoft has been silently fixing bugs as far back as 2004. He referred to the companys MS04-007
bulletin as a classic example of Microsoft announcing a fix for a single vulnerability when in fact a total of seven flaws were quietly fixed.
Read details here about Microsofts decision to use external patch testers.
Manzuiks team presented a research paper on its findings at the Black Hat Briefings in Europe earlier in 2006 to highlight the problems with withholding details on fixes from customers.
"Microsofts customers depend on that information to figure out how to respond to Patch Tuesday. The reality is, system administrators will delay deploying a patch based on the details of the bulletin. When details arent included, he wont install that patch. That is a big problem," Manzuik said.
He said IT departments do not have the skill or resources to reverse-engineer every patch.
"They are simply left in the dark and may ignore a patch that is super-critical to their environment. Meanwhile, the bad guy has spent the time to find out what was silently fixed," Manzuik said, arguing that Microsoft has a responsibility to make sure businesses are fully informed about software changes.
"I dont buy the argument that they are aiding attackers. The attackers are already reverse-engineering the patches. They have the time and resources to find out where the flaw lies. The guy that feels the pain is the system administrator who is in the dark and who cant do his own reverse-engineering," Manzuik said.
Matthew Murphy, the independent researcher who flagged the issue after finding silent fixes in the April batch of patches,
said third-party vendors that incorporate code from Microsoft are also hurt by the lack of full disclosure.
Murphy outlined a recent case where anti-virus vendor Trend Micro got burned by a silent fix
pushed out by Microsoft. That issue revolved around a bug in Visual Studio that was reported to Microsoft in 2002 but remained unfixed for several years.
Microsoft eventually fixed the bug but information was withheld, causing Trend Micro to unwittingly use the vulnerable code in its products, putting its customers at risk of a heap overflow vulnerability that could be used in code execution attacks.
Manzuik also pointed out that businesses rely heavily on host-based IPS technology to secure valuable assets while patches are being tested for deployment.
"Some of these IPS products need information from the software vendor to create signatures. How can you create a signature for a flaw if you dont know the location of the flaw? We have proven that signature-based technology can be bypassed to exploit these silently fixed flaws," he said.
Reavey said businesses should use Microsofts severity rating system to help with patch deployment timetables. "Its important to remember that the best way to be safe and secure is to apply all the updates. We are providing patches for everything. We still recommend a defense-in-depth strategy that includes IPS and IDS [intrusion detection system] technology, but customers should use our severity ratings system and apply the patches," he said.
Check out eWEEK.coms for the latest security news, reviews and analysis. And for insights on security coverage around the Web, take a look at eWEEK.com Security Center Editor Larry Seltzers Weblog.