Demonstration of Vulnerabilities in GSM Security With USRP B200 and Open-Source Penetration Tools
Demonstration of Vulnerabilities in GSM Security With USRP B200 and Open-Source Penetration Tools
net/publication/307544953
CITATIONS READS
14 7,553
4 authors, including:
Khyati P Vachhani
The MathWorks, Inc
14 PUBLICATIONS 88 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Use of ICT in teaching English to Arts and Commerce students of Gujarat View project
All content following this page was uploaded by Khyati P Vachhani on 01 September 2016.
Abstract—This paper showcases the vulnerabilities in the GSM networks are now facing a multitude of security challenges
security architecture through implementation of an active attack including the well-known man-in-the-middle attacks.
at the Um interface. The attack was carried out by taking
advantage of lack of two-way authentication. A rogue GSM
base transceiver system was established using Universal Software Traditionally GSM interception hardware was costly and
Radio Peripheral (USRP) B200 board and OpenBTS. B200 allows custom built. Procurement of such hardware and endmost
relaxation of an external 10 MHz reference signal as opposed to approval of deploying it was at the behest of government
widely used USRP1 and N-series. After establishing rogue BTS, sponsored organizations only. This limited the analysis and
IMSI catch-attack and impersonation of a mobile subscriber testing of GSM infrastructure and its protocols. However, in
to send malicious SMS are executed. Along with OpenBTS,
standalone standard applications - Asterisk and smqueue are recent years, with the advancement of numerous open-source
used for correct routing of messages. The attacks are observed tools and re-configurable architectures like Software Defined
on the TEST network and not the spoofed network so that no Radio (SDR) [4], access to the technology to analyze and
infringement is established on security and privacy of the existing test GSM infrastructure is now open to almost everyone. Free
GSM subscribers. access to the tools and technology has pushed the boundaries
Index Terms—GSM, SDR, USRP, OpenBTS, GNU Radio,
security, authentication, IMSI-catch, man-in-the-middle attack of the attacker too. Newer variants of attacks like one-way
authentication, man-in-the-middle, denial of service etc have
been discovered and reported [5].
I. I NTRODUCTION
Cellular networks have grown continually over the past This paper focuses on examining those vulnerabilities of the
decades and their utilization has increased manifold in both GSM protocol stack that permit us to perform passive/active
developed and developing countries [1]. Cell phone utilization attacks on the network. Man-in-the-middle attacks are imple-
is expected to reach 95 percent of the world population by the mented using Universal Software Radio Peripheral (USRP)
end of 2016. Despite the exponential growth of such networks, hardware [6] and Open Base Transceiver Station (OpenBTS)
one can find the presence of outdated infrastructure in the software [7]. In addition to this, the authors show it is also
network. The foremost reason for not replacing the outdated possible to send defamatory Short Message Service (SMS)
infrastructure is the huge capital requirement for upgradation. to multiple subscribers. Since it is of utmost importance to
Also deployment of latest technology is confined to densely take care of the privacy of other GSM subscribers during the
populated regions because of higher demands and low return attacks, the non-encrypted traffic is acquired from the BTS
on investment [2]. Furthermore, compatibility with various and a separate isolated room is used for testing the attacks
legacy systems and previous standards has become an essential and sending defamatory SMS.
part with latest mobile handsets. Multiband compatibility
renders economic benefits for the mobile network operators The paper is organized as follows: Section 2 briefly de-
(MNOs) but at the cost of sabotaging the security of the scribes the security architecture of the GSM system which is
network. There is no existing deployed technology which has followed by the possible security threats on GSM standards in
manifested impassable security structure. This demands crucial section 3. Section 4 highlights the attack scenario in GSM sys-
steps to be incorporated to provide highly secured services tem using rogue base station. The various hardware−software
and more advanced means to examine existing technologies tools used to establish the attacks are discussed in section 5.
[3]. With attackers having decades of knowledge of several Section 6 explains the experimental setup requirements fol-
vulnerabilities and ever increasing surge of computing power lowed by the observations carried out during the deployment
available, Global System for Mobile Communication (GSM) of the system. Finally, the paper ends with the conclusion.
II. S ECURITY A RCHITECTURE IN GSM rather robust and impenetrable version of the algorithm in such
The security architecture of the GSM standard looks after a way that it does not hamper the efficiency of the network or
provision of security services like authentication, confiden- introduce major variation in the infrastructure. This target can
tiality of the subscriber data and air interface encryption [8]. be achieved if new algorithms follow predefined standard in-
Each of these are extremely crucial for an end-to-end secured terface, i.e., key length. Hence, the A5 encryption algorithm is
communication. The basic security architecture of a GSM deployed, which actually is a lineage of algorithms that realize
network is shown in Figure 1. A brief description of these a predefined standard interface [9]. The standard versions with
security aspects is as follows: 64-bit keys are A5/1, A5/2 and A5/3. A5 version 0 however
has no encryption algorithm, necessitating a warning from the
network to the subscriber about transmission of information
over the air interface in plain-text. This warning message must
be broadcasted on the MS display screen, thus activating the
ciphering indicator feature in the SIM.
on the uplink. Decreasing the downlink transmission power Thus, MS is still under the impression that it has moved into
will further clean up the uplink. The coverage area lost by a new cell circle. This turns out to be an advantage for the
decreasing the downlink power is not significant in a lab attacker in avoiding the MS to connect to other nearby BTS.
environment. Cleaner signals are preferable to strong ones. To avoid MS getting connected to other BTS, the attacker can
With the necessary information of the 2G serving BTS and set a cell reselect hysteresis (CRH) to a high value and increase
the list of neighboring ARFCN, the attacker is now fully the transmit power.
prepared to establish a rogue BTS. 1) IMSI-catch using Rogue Base Station: After establishing
the rogue BTS and setting the necessary environmental param-
eters, the IMSI-catch attack could be successfully executed
in few seconds. Handsets will not usually divulge the IMSI
of their SIM card. It can sometimes be located in a menu
or through a field test mode, but this method of determining
a SIMs IMSI is very cumbersome. Luckily, there are other
methods; OpenBTS also knows the IMSIs it has interacted
with because the attacker is in control of the network side and
also has access to this information. To force an interaction
between a handset and test network, the attacker has to perform
a Location Update Request (LUR) operation on the network,
analogous to a registration. This is nothing more complicated
than selecting the network from the carrier selection list.
The IMSI numbers caught are shown in Figure 7. The pri-
mary idea of this attack is to make use of these IMSI numbers
in connecting with the real network. The attacker will receive
the RAND from the real network which in turn is forwarded to
the attacked mobile user for calculating SRES value. Now, the
attacked mobile user reverts the SRES to the attacker assuming
that the real network initiated the process. Ultimately, the
attacker forwards the SRES value from rogue BTS to the real
network where successful authentication is accomplished. In
this way, the IMSI-catch attack by an attacker becomes a man-
Fig. 5: Step-by-Step Active Attack flow in-the-middle attack between the mobile subscriber and the
real network.
Depending on the handset model, firmware, and SIM used, The AUTH parameter in Figure 8 shows if the MS is
the network ID will be displayed as 00101, 001-01, Test authenticated or not. AUTH 0 means that the location update
PLMN 1-1, as shown in Figure 6. request was rejected. AUTH 1 means that the MS is a
subscriber in the network and is authenticated. AUTH 2 means
that the MS is not a subscriber, but the location update request
is accepted. In this case, the MS is camped on the network,
but does not have a phone number. One can further observe
that all except one of the caught IMSIs had a AUTH value
of 2. The one with AUTH value of 1 is an MS that was
added as a subscriber to the network before the experiment
was conducted.
The attacker will now trigger the rogue BTS through cell Fig. 7: list of IMSI numbers caught
reselection process. This process takes place mainly due to
two reasons : 2) Sending malicious SMS using Rogue Base Station: In
(1) the attacker is using one of the neighboring cell ARFCNs addition to IMSI-catch attacks, an attacker can also imperson-
(2) the transmission power of the rogue BTS is higher than ate a mobile user to send malicious SMS messages from the
the serving BTS. rogue BTS. It is quite critical to observe that it is equally