Huge Web Hack Attack Infects 500,000 Pages

Attacks on legitimate Web domains, including some belonging to the United Nations that began earlier this week, have expanded dramatically, security researchers said Friday, with hundreds of thousands of pages hacked by Friday.

One anti-virus vendor said the sites might have been compromised through a "security issue" in Microsoft's Web server software that has been reported to Microsoft's engineers.

On Wednesday, several security companies, including California-based Websense, said large numbers of legitimate sites, including URLs for the U.N., had been hacked and were serving up malware. These latest site compromises were only the most recent SQL injection attacks, however; similar attacks have been launched since the first of the year, and were last detected in large numbers in March.

Earlier in the week, Dan Hubbard, Websense's vice president of security research, estimated the number of hacked sites in the low six figures. By today, that number had soared as firms such as Panda Security pegged the number at 282,000, and F-Secure said its infected-page count was above half a million.

Ryan Sherstobitoff, a corporate evangelist for Panda, said his company had reported a problem with Internet Information Services (ISS) to Microsoft that was probably responsible for the hacks. "We reported a security issue, but I don't have any specific details on whether it's a vulnerability," Sherstobitoff said.

"It's not like this is a brand-new problem," he said, referring to legitimate site compromises. "But Microsoft has already issued a security advisory that said they are investigating public reports of problems with IIS. This seems to be related to that advisory."

That advisory was published April 17, and warned users of a bug in most versions of Windows that could be exploited through custom Web applications running in IIS. It could also be exploited via SQL Server, Microsoft said.

On Friday, Microsoft said it did not know whether the ongoing site attacks were linked to the bug described in the April 17 advisory. "We have not yet determined whether or not these reports are related to Microsoft Security Advisory 951306 released last week," a company spokesman said in an e-mail.

Microsoft also contested Panda's claim that it had reported a problem. "Microsoft is currently aware of and is reviewing reports regarding public claims of attacks on IIS Web servers," said Bill Sisk, a communications manager who works in the Microsoft Security Response Center. "While we have not been contacted directly regarding these reports, we will continue to monitor all reports either publically shared or responsibly disclosed and investigate once sufficient details are provided."

How It Works

Although it may not be clear how attackers are compromising such large numbers of Web sites, what happens after a site is infected is well-understood, researchers have said. When a visitor reaches one of the hacked sites, malicious JavaScript loads an IFRAME from a malware-hosting server; the IFRAME redirects the browser to a different page, also hosted on the hacker's server.

Next, a multiple-strike attack kit is downloaded to the visitor's PC. The kit tries eight different exploits, and if it finds one that works, hijacks the system.

These kinds of attacks, said Sherstobitoff, essentially make the idea of a "trusted site" moot. "You used to know that if you walked down the dark streets of the Web, you would be infected. Today, you really can't tell what the dark streets are."

The hacker strategy, of course, is to leverage that uncertainty. "This is getting really bad," Sherstobitoff said.

It's so bad, in fact, that while security companies urged Web site administrators to check their server logs for evidence of a compromise, and told corporate security staffs to block several malware-hosting sites at their companies' perimeters, they didn't have much useful advice for end-users.

"Users should be extremely wary when visiting sites, even those typically trusted," was about all Symantec could come up with in an alert to customers of its DeepSight threat notification service.

Disabling JavaScript can also protect against such attacks, Symantec added. Users, however, are often reluctant to switch off JavaScript because without it, many sites are crippled or won't display properly.