Demons Lurk in Management Software

You know that management agent software that sits on nearly every server and workstation in your network? It may be the most dangerous software in your enterprise.

Researchers from Matasano Security have discovered holes in systems management software that allow it to be used for mass attacks within the network, either via an infected central management server that would take control of thousands of machines running its agent component, or via an infected client machine that could poison the management server. Matasano researchers will report their findings on these vulnerabilities at the upcoming Black Hat USA conference next month in Las Vegas.

Tom Ptacek, a researcher for Matasano, a security consulting firm, says during the past few months he and his colleagues discovered disturbing vulnerabilities in systems management software while testing out their thesis that non-Internet applications are ripe for attack. Ptacek wouldn't name any specific management products that fell prey to his firm's staged attacks, but he says they included major systems management vendors that handle things like patch management. This represents a $2 billion to $3 billion market, he says.

The researchers were able to crack the encryption algorithms around vendor-specific protocols and use their homegrown agents and "evil" management systems to see how much damage they could do. One glaring problem: The management server was sure to "call" the agent system to check on its software update status, Ptacek says, so it was easy for an infected agent to spread its poison to the server. Then the management server becomes a stepping stone to controlling all other machines, either by installing rootkits on them or grabbing sensitive information.

"We worry about an attacker seizing machines and owning them perpetually," says Ptacek, who was careful not to give away too much detail before management vendors get a chance to issue patches. "What they could do with access to thousands of machines is pretty scary."

These types of attacks so far haven't registered on the radar screen. And Matasano has been keeping management vendors apprised of the problems they found. "We don't know of anyone using these specific vulnerabilities," says Window Snyder, CTO of Matasano. "But this type of attack is so obvious it must have occurred."

And there's no telling what the overall impact would be of such an attack, Snyder says. But she's sure it would definitely be more widespread than an Internet-borne attack on a DMZ.

So just how serious is this systems management agent threat? "A seven on a one to 10 scale," says Richard Stiennon, chief research analyst for IT-Harvest. Stiennon says security management consoles could be next on the list.

Enterprises today are disciplined about patching and updating their operating systems and other software, says Danny McPherson, chief research officer for Arbor Networks. "But how vulnerable are the platforms that patch them?"

The threat to internal applications has been largely ignored to date, with security technology mostly focused on the perimeter. The bottom line is it's tough to build secure software applications, Ptacek says. But it seems to be slowly getting some much-needed attention -- startup V.i. Laboratories, which launched today, is honing in on application security. (See Startup Locks Down Apps.) And just last week, IBM announced tools for building security around applications.

Matasano researchers, meanwhile, had been probing various internal applications over the past year for vulnerabilities. They found the first big hole in an iSCSI storage appliance, which researchers were able to easily penetrate and grab its data without even having the password, Ptacek says. But systems management software had more far-reaching security risks.

With internal VPNs and wireless blurring the lines of the internal network, the inside is more susceptible than ever, says Arbor Networks' McPherson. "The target is more the inside" now, he says, so companies have to know their network- and host-level vulnerabilities.

"We don't want to wait until we're relying on the security of software to prevent attackers from wreaking havoc on enterprises," says Ptacek. "Right now, internal applications are not secure enough to withstand an internal attack. So it's time to take a very serious look."

— Kelly Jackson Higgins, Senior Editor, Dark Reading

Companies mentioned in this article: