Enterprises that use mobile device management (MDM) systems to protect their corporate data on employees' mobile phones are not safe from attacks from spyphones, researchers warned at BlackHat Europe on Thursday.
Over the next five years, 65 percent of enterprises will adopt a mobile device management (MDM) system for corporate users, technology research company Gartner predicted last October. Companies will use the systems to manage network traffic and corporate data on smartphones and tablets, which nowadays are often owned by employees and used for both private and corporate tasks.
Companies are using MDM systems to protect their data, but they must be aware that while the systems are useful, they don't provide full security and can be targeted by so-called spyphones, warned Daniel Brodie, senior security researcher at the Israeli security company Lacoon Security, and Michael Shaulov, CEO and co-founder of the company, at the BlackHat conference in Amsterdam.
MDM systems try to tackle security issues by providing a "secure container" on mobile devices, encrypting the part of the mobile device that handles business data, as well as offering the possibility to remotely wipe or lock that section if a phone is stolen or an employee quits. However, common MDM security offerings can be circumvented by planting surveillance tools without the users knowledge on a phone, turning it into a spyphone, Brodie told a crowd of conference attendees.
A survey conducted by Lacoon in cooperation with global cellular network providers showed that about one in 1,000 phones was a spyphone, according to Brodie's research paper. Of 175 compromised devices found, 52 percent was attributed to Apple's iOS, 35 percent to Android phones, 7 percent to Nokia phones and 6 percent to other devices, he said.
"This is a very alarming number," Brodie said. The problem with spyphones is that while the software is installed on a single device, it is used to target whole organisations for espionage purposes, Brodie said. And as such, the impact of a spyphone attack on an organisation can be "extremely high," he added.
Most spyphones are used for recording confidential phone calls and board meetings, tracking locations, extracting call logs as well as text messages and voice memos, and snooping on corporate emails and application data, Brodie said.
Secure containers of MDM systems can be bypassed in order to install spyphone software. On Android devices this can be done by publishing a seemingly innocent application in an Android market. Once the victim has installed the app, the app refers to the malicious code, which is then downloaded, the researchers said. After this, the spyphone creates a hidden binary and uses it for privileged operations, such as reading mobile logs.
iOS devices are much harder to crack but are probably more appealing to spyphone makers since a lot of companies are standardising on iOS, Brodie said. An attacker has to install a signed application on the targeted device using an enterprise- developer certificate. The attacker then uses a jailbreak exploit -- removing limitations and protections to gain root access to iOS -- to inject container-bypass code into the secure container. After that, the attacker removes every source of the jailbreak.
"If you're looking at the phone you're not going to see if it's jailbroken," said Brodie, adding that he and Shalouv had several times tried to jailbreak an iPhone that was already jailbroken. They simply did not notice it already was, they said.
Once the jailbreak is removed, the spyphone places hooks in the secure container using Objective-C hooking mechanisms. The spyphone is than alerted when an email is read, is able to pull the email and subsequently sends every loaded email to a command and control (C&C) server that is controlled by the attackers, according to Brodie.
While mobile OSes try to protect themselves by protecting the OS from attackers and users, jailbreaking and rooting methods are rendering this security mechanism irrelevant, according to Brodie.
"Infection is inevitable," Brodie said. This however doesn't mean that MDMs are not useful. They are useful for separating personal and business data and also can be very useful when use for remote-wipe operations, the researchers said.
Companies need to be aware though that MDMs cannot provide absolute security, the researchers said. The security industry therefore should try to find a way to solve this problem, they added. Solutions could for instance look at different network parameters and abnormal behavior to signify an infected device. Those parameters for example could consist of behavioral analysis to signify strange behavior, traffic to well known C&C servers, and data intrusion detection, they said.
Loek is Amsterdam Correspondent and covers online privacy, intellectual property, open-source and online payment issues for the IDG News Service. Follow him on Twitter at @loekessers or email tips and comments to [email protected]