By Parmy Olson
Security researcher Karsten Nohl says some
SIM cards can be compromised because of wrongly configured Java Card
software and weak encryption keys; Photo credit Luca Melette
Smartphones are
susceptible to malware and carriers have
enabled NSA
snooping, but the prevailing wisdom has it there’s still one part of
your mobile phone that remains safe and un-hackable: your SIM card.
Yet after three years of research, German cryptographer Karsten Nohl
claims to have finally found encryption and software flaws that could
affect millions of SIM cards, and open up another route on mobile phones
for surveillance and fraud.
Nohl, who will be
presenting his findings
at the Black Hat security conference in Las Vegas on July 31, says his
is the first hack of its kind in a decade, and comes after he and his
team tested close to 1,000 SIM cards for vulnerabilities, exploited by
simply sending a hidden SMS. The two-part flaw, based on an old security
standard and badly configured code, could allow hackers to remotely
infect a SIM with a virus that sends premium text messages (draining a
mobile phone bill), surreptitiously re-direct and record calls, and —
with the right combination of bugs — carry out payment system fraud.
Payment fraud could be a particular problem for mobile phone users in
Africa, where SIM-card based payments are widespread. The deployment of
so-called NFC payment technology, already
slow to take off, could also be at risk, Nohl says, as well as the ability for carriers to track charges to each caller’s account.
There’s no obvious pattern to the flaw beyond the premise of an
older encryption standard. “Different shipments of SIM cards either have
[the bug] or not,” says Nohl, who is chief scientist at risk management
firm Security Research Labs. “It’s very random.”
In his study, Nohl says just under a quarter of all the SIM cards he
tested could be hacked, but given that encryption standards vary widely
between countries, he estimates an eighth of the world’s SIM cards could
be vulnerable, or about half a billion mobile devices.
Nohl, who was
profiled by
Forbes’ Andy
Greenberg in 2011 for his work on breaking mobile encryption standards,
believes it unlikely that cyber criminals have already found the bug.
Now that word of the vulnerability is out, he expects it would take them
at least six months to crack it, by which time the wireless
industry will have implemented available fixes.
That effort may already be underway. Nohl says at least two large
carriers have already tasked their staff with finding a patch for the
SIM vulnerability, which they will share with other operators through
the wireless trade body GSMA.
“Companies are surprisingly open to the idea of working
cooperatively on security topics because the competition is somewhere
else,” says Nohl. “The competition is organized crime,
not AT&T versus T-Mobile.” (The situation in similarly in finance,
where payment services like MasterCard, Visa, and American Express will
work together under industry association
EMVco to improve security standards for smart cards.)
The market for SIMs is almost entirely fed by mobile carriers, and
supplied by two leading global vendors, Gemalto and Oberthur
Technologies. Both have profited heavily from the huge growth in mobile
handsets: two years ago there were 1 billion SIM cards worldwide, and
today there are more than 5 billion, says ABI Research analyst John
Devlin, though the market is slowly reaching a plateau. SIMs are thought
to be one of the most secure parts of a phone, he added, and as the
carrier’s property, are “key to their relationship between you and I,
the subscriber.”
Vodafone would not answer questions about the level of encryption
its SIM cards used, and referred all media questions to GSMA. Both
Verizon and AT&T said they knew of Nohl’s research, but said their
SIM profiles were not vulnerable to the flaw. AT&T added that it had
used SIMs with triple Data Encryption Standards (3DES) for almost a
decade; Verizon did not specify why its SIMs were not vulnerable.
The London-based GSMA said it had looked at Nohl’s analysis and
concurred that “a minority of SIMs produced against older standards
could be vulnerable.” It said it had already provided guidance to
network operators and SIM vendors who could be impacted by the flaw.
“There is no evidence to suggest that today’s more secure SIMs, which
are used to support a range of advanced services, will be affected,” a
spokesperson added.
Nohl says that while AT&T and Verizon may benefit from robust
SIM encryption standards, other carriers will use straight Data
Encryption Standards (DES), guidelines developed in the 1970s that are
fundamental to why he was able to “get root” on dozens of SIMs cards.
“Give me any phone number and there is some chance I will, a few
minutes later, be able to remotely control this SIM card and even make a
copy of it,” Nohl says.
SIM cards are essentially mini-computers with their own operating
system and pre-installed software. To maintain security, many rely on a
cryptographic standard called
DES (digital encryption standard),
which was invented by IBM in the 1970s and improved by the NSA. Some
networks, like AT&T and the four major carriers in Germany, have
moved away from using the old version of the standard, but others have
not. Though Nohl didn’t identify a pattern to vulnerable SIMs in terms
of manufacturers, the ones he could hack all used the old encryption
standard.
Key to the hack is
Java Card,
a general purpose programming language used on 6 billion SIM cards. If
operators need to update something on your SIM, for instance allowing
interoperability with a carrier in another country, it will execute the
right Java Card programs on your SIM by sending your mobile a binary
SMS. This is a text message you will never see, sent through a method
called over-the-air programming (OTA).
In early 2011, Nohl’s team started toying with the OTA protocol
and noticed that when they used it to send commands to several SIM
cards, some would refuse the command due to an incorrect cryptographic
signature, while a few of those would also put a cryptographic signature
on this error message.
With that signature and using a well known cryptographic method
called rainbow tables, Nohl was able to crack the encryption key on the
SIM card in about one minute. Carriers use this key to remotely program
a SIM, and it is unique to each card.
“Anybody who learns the key of a particular SIM can load any
application on the SIM he wants, including malicious code,” says Jasper
Van Woudenberg, CTO North America of smart-card security firm Riscure.
“We had almost given up on the idea of breaking the most widely
deployed use of standard cryptography,” says Nohl, but it felt “great”
to finally gain control of a SIM after many months of unsuccessful
testing.
With the all-important (and till-now elusive) encryption key, Nohl
could download a virus onto the SIM card that could send premium text
messages, collect location data, make premium calls or re-route calls. A
malicious hacker could eavesdrop on calls, albeit with the SIM owner
probably noticing some suspiciously-slow connections.
Nohl found a second bug. Unrelated to the weak encryption key, it
allows even deeper hacking on SIMs and is caused, Nohl says, by a
mistake on the part of SIM card manufacturers. Java Card uses a concept
called sandboxing, in which pre-installed programs like a Visa or PayPal
app are shielded from one another and the rest of the SIM card. The
term comes from the idea of only allowing programs to “play with their
own toys, in their own sandbox,” says Nohl. “This sandboxing mechanism
is broken in the most widely-used SIM cards.” The researcher says he
found a few instances where the protocols on the SIM card allowed the
virus he had sent to a SIM, to check the files of a payment app that was
also installed on the card.
The way this works is somewhat complex, but Nohl’s
virus essentially gave the infected Java software a command it could not
understand or complete – eg. asking for the 12th item in a 10-item
list, leading the software to forgo basic security checks and granting
the virus full memory access, or “root,” in cyber security parlance.
In sum, a malicious hacker who wanted to use this method might
start with a list of 100 phones. They could send a binary SMS to all of
them, using a programmable cell phone connected to a computer. They
might get 25 responses with cryptographic signatures, and dismiss the
half that use a stronger security standard. From the rest, Nohl surmises
they could crack the encryption key of perhaps 13 SIM cards, and send
them a virus that breaks through the Java Card sandbox barriers and
reads payment app details, as well as the master key of the SIM card.
Who’s to blame for this and who can fix it? Nohl says broken Java
sandboxing is a shortcoming of leading SIM card vendors like Gemalto and
Oberthur. Riscure’s Van Woudenberg agrees.
Gemalto which
made about half its
$2.5 billion revenue in 2012 selling SIM cards, said in an email to
Forbes that
its SIMs were “consistent with state-of-the-art and applicable security
guidelines,” and that it had been working closely with GSMA and other
industry bodies to look into Nohl’s research. Gemalto’s CEO Olivier Piou
has
said publicly that there are no security issues with mobile payments, and his company
says on its website that SIM cards are “virtually impossible to crack.”
Despite this, Nohl believes badly-configured Java Card sandboxing
“affects every operator who uses cards from two main vendors,” including
carriers like AT&T and Verizon who use robust encryption standards.
Are SIM cards with these 3DES standards vulnerable? Nohl suggests they
might be, and that he’ll expound on the details at Black Hat.
At minimum it seems that carriers should upgrade to newer encryptions
quickly, not just for the safety of their subscribers, but future
revenue too. Payment providers like MasterCard and Visa will need to use the OTA protocol to
fill SIM cards with Java applications, like credit card applets, and
enable NFC-based payments on phones in the future — and they’ll pay
carriers for the privilege of being on the SIM. “Operators see this as
valuable real estate,” says Nohl, referring to this OTA communication
channel. Leaving aside what this means for consumer privacy, Nohl’s
findings may leave some carriers grappling with new questions over the
security (and value) of this real estate.
“Carriers and SIM card manufacturers do need to step up their
security game for when payments arrive,” says Van Woudenberg. Banks are
slow and cautious with new technology as they wait for it be proven
secure, he adds, but “the mobile world moves much faster, as
time-to-market is for them more important.”
As mobile payments bring these two worlds together, Nohl’s
research has shown the process of proving out security on SIMs could be
more challenging than the key players originally thought.