[go: up one dir, main page]

By Bob Sullivan on Technolog

  • Exclusive: Millions of printers open to devastating hack attack, researchers say

    Columbia University

    This time-lapsed image of a screen on an HP LaserJet shows the impact of a rogue print job used to reprogram the device.

    Could a hacker from half-way around the planet control your printer and give it instructions so frantic that it could eventually catch fire? Or use a hijacked printer as a copy machine for criminals, making it easy to commit identity theft or even take control of entire networks that would otherwise be secure?

    It’s not only possible, but likely, say researchers at Columbia University, who claim they've discovered a new class of computer security flaws that could impact millions of businesses, consumers, and even government agencies.

    Printers can be remotely controlled by computer criminals over the Internet, with the potential to steal personal information, attack otherwise secure networks and even cause physical damage, the researchers argue in a vulnerability warning first reported by msnbc.com.  They say there's no easy fix for the flaw they’ve identified in some Hewlett-Packard LaserJet printer lines – and perhaps on other firms’ printers, too – and there's no way to tell if hackers have already exploited it.


    The researchers, who have working quietly for months in an electronics lab under a series of government and industry grants, described the flaw in a private briefing for federal agencies two weeks ago. They told Hewlett-Packard about it last week.

    HP said Monday that it is still reviewing details of the vulnerability, and is unable to confirm or deny many of the researchers’ claims, but generally disputes the researchers’ characterization of the flaw as widespread.  Keith Moore, chief technologist for HP's printer division, said the firm "takes this very seriously,” but his initial research suggests the likelihood that the vulnerability can be exploited in the real world is low in most cases.      

    “Until we verify the security issue, it is difficult to comment,” he said, adding that the firm cannot say yet what printer models are impacted.

    But the Columbia researchers say the security vulnerability is so fundamental that it may impact tens of millions of printers and other hardware that use hard-to-update “firmware” that’s flawed.

    'Crystal clear'
    The flaw involves firmware that runs so-called "embedded systems" such as computer printers, which increasingly are packed with functions that make them operate more like full-fledged computers. They also are commonly connected to the Internet. 

    "The problem is, technology companies aren't really looking into this corner of the Internet. But we are," said Columbia professor Salvatore Stolfo, who directed the research in the Computer Science Department of Columbia University’s School of Engineering and Applied Science. “The research on this is crystal clear.  The impact of this is very large. These devices are completely open and available to be exploited.”

    Printer security flaws have long been theorized, but the Columbia researchers say they've discovered the first-ever doorway into millions of printers worldwide.  In one demonstration of an attack based on the flaw, Stolfo and fellow researcher Ang Cui showed how a hijacked computer could be given instructions that would continuously heat up the printer’s fuser – which is designed to dry the ink once it’s applied to paper –  eventually causing the paper to turn brown and smoke.

    In that demonstration, a thermal switch shut the printer down – basically, causing it to self-destruct – before a fire started, but the researchers believe other printers might be used as fire starters, giving computer hackers a dangerous new tool that could allow simple computer code to wreak real-world havoc.

    Hewlett Packard, in a statement, said all its printers include such thermal switches, and these would prevent a printer fire in all cases.

    "(The thermal breaker) cannot be overcome by a firmware change or this proposed vulnerability," it said.

    Click here to read H-P's full statement issued in response to this story.

    Cui and Stolfo say they've reverse engineered software that controls common Hewlett-Packard LaserJet printers. Those printers allow firmware upgrades through a process called "Remote Firmware Update." Every time the printer accepts a job, it checks to see if a software update is included in that job.  But they say printers they examined don't discriminate the source of the update software – a typical digital signature is not used to verify the upgrade software’s authenticity – so anyone can instruct the printer to erase its operating software and install a booby-trapped version.

    In all cases, the Columbia researchers claim, duping a would-be target into printing a virus-laden document is enough to take control of that person's printer; but in some cases, printers are configured to accept print jobs via the Internet, meaning the virus can be installed remotely, without any interaction by the printer's owner.

    “It's like selling a car without selling the keys to lock it,” Stolfo said. “It’s totally insecure.”

    Columbia University

    Columbia researcher Ang Cui explains how he was able to infect an HP printer with malicious code.

    Rewriting the printer's firmware takes only about 30 seconds, and a virus would be virtually impossible to detect once installed. Only pulling the computer chips out of the printer and testing them would reveal an attack, Cui said.  No modern antivirus software has the ability to scan, let alone fix, the software which runs on embedded chips in a printer.

    “First of all, how the hell doesn't HP have a signature or certificate indicating that new firmware is real firmware from HP?” said Mikko Hypponen, head of research at security firm F-Secure, when told of the flaw. “Printers have been a weak spot for many corporate networks.  Many people don’t realize that a  printer is just another computer on a network with exactly the same problems and, if compromised, the same impact.”

    There are plenty of points of contention between HP and the researchers, however. Moore, the HP executive, said the firm’s newer printers do require digitally signed firmware upgrades, and have since 2009. The printers tested by the researchers are older models, Moore said. 

    In contrast, the Columbia researchers say they purchased one of the printers they hacked in September at a major New York City office supply store.

    Moore also said that the impact of any potential vulnerability is limited because most home users have InkJet printers – not LaserJet printers – and they do not permit remote firmware upgrade, he said.

    Still, a widespread flaw in LaserJet printers would raise serious issues. Hewlett Packard dominates the printer market; the firm says it's sold 100 million LaserJet printers since 1984, meaning millions of computers could be vulnerable. HP, by far the dominant printer seller worldwide with 42 percent of the market, sells about 50 million printers of all kinds annually, according to IDC.

    In an exclusive demonstration for msnbc.com at Columbia University’s Intrusion Detection Systems Laboratory, Cui and Stolfo revealed the kind of havoc an attacker could wreak once they gained control of a printer. After sending a virus-laced print job to a target printer, the device's small screen read, in sequence, "Erasing...Programming...Code Update Complete."

    In one demonstration, Cui printed a tax return on an infected printer, which in turn sent the tax form to a second computer playing the part of a hacker’s machine. The latter computer then scanned the document for critical information such as Social Security numbers, and when it found one, automatically published it on a Twitter feed.

    A hacker who merely wanted to wreak havoc could easily disable thousands – or perhaps millions – of vulnerable printers, Cui said, as it is trivial to send the printer upgrades that would render it inoperable.  

    Beachhead?
    But the researchers say the possibilities created by hijacked printers go far beyond pranks or identity theft. Printers on a company network are nearly always trusted by other computers. A hijacked printer could act as a beachhead to attack a company's network that was otherwise protected by a firewall. Few companies are prepared to protect themselves from an attack by their own printer.

    Moore also disagreed with this assertion. He said standard print jobs could not be used to initiate a firmware upgrade; only specially-crafted files sent directly to the printer can do that. Were that true, the vulnerability could only be exploited on printers left exposed to the Internet; printers behind a firewall would be safe.

    “This (vulnerability) is probably not as broad as what I had heard in their first announcement,” Moore said. “It sounds like we disagree on what the exposure might be.”

    But the Columbia researchers say standard print commands sent both from a Macintosh computer and a PC running Linux tricked an HP printer into reprogramming itself. Moore later conceded that might be true; but the two sides disagreed on whether users in a Microsoft Windows environment were safe from the attack.

    Even home users with printers that are not directly connected to the Internet are at risk, Cui said.  As long as the printer is connected to a computer – through a USB cable, for example –  it could be used to launch attacks, or as part of a botnet.

    A quick scan of unprotected printers left open to Internet attack by the researchers found 40,000 devices that they said could be infected within minutes. 

    Cui discovered the lack of authentication by physically disassembling the printer, and painstakingly reading output from its chipset, one character at a time. The chips run off-the-shelf operating systems like VxWorks and Linx, a scaled-down version of the Linux operating system designed for embedded devices.  Reprogramming the chip was relatively easy, he said – and now that the concept has been proven, he thinks others could reproduce his work in a day or two. 

    "In fact, it's almost impossible to think that someone else hasn't already done this," he said.

    Fixing the flaw will not be easy, Stolfo said.  There is no natural path to update printer operating system software, as there is for desktop PC software.  It's possible a consortium of firms could "push out a fix," once one is available, he said. He urged HP to work with companies like Microsoft to help consumers update their printers. (Msnbc.com is a joint venture of Microsoft and NBC Universal.)

    One particularly vexing part of the fix: Printers that are already compromised by rogue software likely cannot be fixed. An attacker could easily shut down the pathway for future updates that would “cure” an infected printer.

    “If and when HP rolls out a fix, if a printer is already compromised, the fix would be completely ineffective.  Once you own the firmware, you own it forever. That’s why this problem is so serious, and so different,” Cui said. “This is nothing like fixing a virus on your PC.”

    Such inability to help consumers manually secure their printers could ultimately have disastrous consequences, Stolfo said.

     “It may ultimately lead to telling everyone they just have to throw their printers out and start over,” he said. "Fixing this is going to require a very coordinated effort by the industry," Stolfo said.

    Rogue software
    Hypponen said that the anti-virus industry could develop software tools that would detect booby-trapped print jobs in word processing documents or emails, and thwart attempts to update printers with rogue software that way. But such an approach would hardly be foolproof.

    The Columbia researchers are just beginning to sample printers sold by other manufacturers; the research is inconclusive so far, but Stolfo and Cui believe the problem is not limited to Hewlett-Packard machines.

     “I think it is very wise to broadcast the problem as soon as possible so all of the printer manufacturers start looking at their security architectures more seriously,” Stolfo said.  “It is conceivable that all printers are vulnerable. …Printers that are 3-, 4-, 5-years-old and older, I’d think, all used unsigned software. The question is, ‘How many of those printers are out there?’ It could be much more than 100 million.”

    That’s why Stolfo and Cui decided to go public with the vulnerability: They believe the sheer scope of the flaw requires immediate attention and cooperation from multiple elements of the tech industry. The two are currently helping HP devise a mitigation strategy.

    HP continues to research the potential flaw, but it’s too early for the firm to announce which products might be impacted, or what consumers should do.

    “Until we know things like whether Windows users are affected, whether this is a class or specific product issue, it is frankly irresponsible to say more,” Moore said.  “If this turns out to be the broad (problem) that's being discussed…we will reach out to customers and get it fixed.  We support our customers and value their trust.”

    Printers, however, are just the tip of the iceberg when it comes to vulnerable embedded devices, Stolfo warned.  Columbia researchers have found that many gadgets now wired to connect to the Internet – including DVD players, telephone conference tools, even home appliances – have no security at all.

    "Right now, very few people are thinking about the security of all these devices, so we're moving on to look at many more of them,” Stolfo said, noting that supposedly secure offices – even in sensitive government agencies – have networked teleconferencing devices, printers, even thermostats that create security risks.

    “This is a whole area that is being ignored,” he continued. “While most folks are focused on applications, there is a comfort level with (embedded systems) that is nonsensical. There's no focus on the security of these devices we take for granted and we carry into secure environments every day.”

    Don't miss the next Red Tape:
    *Get Red Tape headlines on your Facebook Wall
    *Follow Bob on Twitter. 
    *Get an e-mail newsletter with Red Tape stories (requires Newsvine registration).

  • After 16 hours on air at Wall Street protests, a Ustream star is born

    Bob Sullivan / msnbc.com

    Tim Pool at Zuccotti Park.

    The revolution will be Ustreamed. 

    When police ran Occupy Wall Street protestors out of Zuccotti Park in the middle of the night on Tuesday, there was really only one way to watch it live: On Tim Pool's "TheOther99" video stream. Pool, armed only with a cellphone and donated backup batteries, filmed the event through the night. He hosted the coverage news anchor style, the way Brian Williams would, talking almost continuously, stopping occasionally to conduct interviews.

    It might have looked a bit like grainy home video, but Pool had a sizable international audience. His Ustream.TV "channel" had, at various times, more than 20,000 simultaneous viewers, an audience some cable channels might envy. The audience exploded when word went out across Twitter that Pool’s stream was the best way to watch the protests online.

    When I met Tim Tuesday afternoon, he'd been "on the air" continually for nearly 16 hours. And while plenty of video streams have come and gone during the protests, Pool’s broadcast earned him a lot of credibility with protesters, and he was still going strong into the evening.

    "Other than my hand cramping up from holding up the phone, I feel pretty good," he said.  He hadn't gone dark, or gone to the bathroom, that entire time.  "I do really need a piece of fruit, though," he added.

    Pool, from Chicago, has been at the Occupy Wall Street protest from the start. He said had no experience hosting a TV show or live stream, but honed his ability to fill air time with talk during the past two months while documenting the protest. Anyone -- including police officers and foul-mouthed protesters -- who walked near Pool risked being broadcast instantly to the world. 

    Pool spent most of the time Tuesday morning running after crowds of evicted protesters as they tried to reorganize, or showing live video of sanitation workers gathering tents and other personal items in the park for disposal.  He rarely turned the camera on himself.

    "I'm here to document what's going on," he said. "I've been doing this since the beginning."

    Bob Sullivan / msnbc.com

    Tim Pool uses simple equipment to "broadcast" live video of the protest on the Internet.

    What Pool’s doing is vaguely similar to what’s called “lifecasting,” where individuals chronicle their lives online through a continual video stream or similar real-time techniques.  The difference is this: Pool’s all-night broadcast on Tuesday morning was riveting.

    The quality of his video stream -- both in content and technology -- is surprisingly good. And the simplicity of his gear can't be beat. He's using a Samsung Galaxy S2 on Sprint's 4G network to stream video, using the onboard camera and microphone to record, and connecting the phone to a small but powerful backup battery.  If you want to know, it’s an "Energi to Go" 18,000 battery produced by Energizer which provides 18,000 milliamp hours (roughly 10 cell phone charging cycles). When his backup battery drained dangerously low, he put out a plea for help on his stream and received two more donated batteries.

    "That's why I've been able to keep doing this long," he said.

    Perhaps the most complicated part of his video gear was the umbrella he had to raise to protect his gear Tuesday night when rain began falling in lower Manhattan. But by that time, police had allowed protesters back into Zuccotti Park, and Pool wasn't about to let a little rain get in the way of his broadcast.

    Don't miss the next Red Tape:
    *Get Red Tape headlines on your Facebook Wall
    *Follow Bob on Twitter. 
    *Get an e-mail newsletter with Red Tape stories (requires Newsvine registration).

  • Siri lets strangers control some iPhone functions without a password

    Rosa Golijan / msnbc.com

    Msnbc.com's Rosa Golijan was able to send a message with a locked phone.

    Apple's new Siri iPhone 4S personal assistant may be a lot of fun, but she's far too willing to talk with strangers, says U.K.-based security research firm Sophos Ltd.

    There's been plenty of talk about Siri's funny personality quirks, and how she will respond with sharp wit -- even when owners swear at her or talk dirty to her.

    The problem is this same sense of wit wasn't applied to Siri's security settings, according to Sophos.


    By default, Siri will take commands from anyone, even if the phone itself is in lock-down mode,  researcher Graham Cluleysaid in a blog post. That means a stranger could pick up a locked iPhone 4S and send an e-mail, send a text message or many other things the rightful owner of the gadget can do -- without needing to enter a four-digit passcode.

    As software developers are fond of saying, this isn't a bug, it's a feature.

    Users who delve into their iPhone security settings will find an option, turned on by default, that says "allow access to Siri when locked with a passcode."

    That's a poor security design choice, Cluley said.

    "What's disappointing to me though is that Apple had a clear choice here," he said. "They could have chosen to implement Siri securely, but instead they decided to default to a mode which is more about impressing your buddies than securing your calendar and e-mail system," he wrote.

    Apple didn’t immediately respond to requests for comment.

    Tests at msnbc.com show Siri’s friendliness doesn’t extend to all iPhone features. She wouldn’t read out existing e-mails without prompting users for a passcode.

    Rosa Golijan / msnbc.com

    Siri coughed up contact information, too, without demanding a password.

    But msnbc.com's Rosa Golijan found a long list of iPhone features she could perform, even while the phone was locked. She was able to: look up contacts and retrieve full contact information, including nicknames; learn who sent text messages to the owner; make phone calls, including international calls; and turn alarms on and off.

    Curiously, Siri refused to do other things: she wouldn't reveal recent call history, for example, or spit out directions. And most critically, when asked to unlock the phone, Siri said, "I'm sorry, I can't do that." 

    The selective behavior means Apple considered security when creating Siri's locked-phone permissions.

    "But there's plenty you can do with the phone locked, and there's a lot of damage you can cause," Golijan said.

    Users concerned about Siri being overly friendly can turn off the Siri security bypass in the Settings/General/Passcode Lock menu on their iPhones.

    "Those of us who work in the security arena have often banged on about the importance of securing your smartphone with a password or passcode to prevent unauthorized access," Cluley said. "Most mobile phone manufacturers have recognized that as so many people use their smartphones to manage their diaries, their private communications and their social lives, it's good to have some form of security."

    With reporting by Rosa Golijan.

    Follow Bob Sullivan on Facebook or Twitter. 

    We take a moment to chat with the iPhone 4S's Siri personal servant, and find that she's not only very helpful and surprisingly accurate, she knows how to joke around and tell off smart alecks, too.