How to Hack into Facebook without being a Hacker
Abhay Patel
Jetking
Tarun Parwani
Rutgers University
tarun.parwani@rutgers.edu
Ramin Kholoussi
Rutgers University
rk496@rutgers.edu
Panagiotis Karras
Rutgers University
karras@business.rutgers.edu
ABSTRACT
The proliferation of online social networking services has
aroused privacy concerns among the general public. The
focus of such concerns has typically revolved around providing
explicit privacy guarantees to users and letting users
take control of the privacy-threatening aspects of their online
behavior, so as to ensure that private personal information
and materials are not made available to other parties and
not used for unintended purposes without the user’s consent.
As such protective features are usually opt-in, users have to
explicitly opt-in for them in order to avoid compromising
their privacy. Besides, third-party applications may acquire
a user’s personal information, but only after they have been
granted consent by the user. If we also consider potential
network security attacks that intercept or misdirect a user’s
online communication, it would appear that the discussion
of user vulnerability has accurately delimited the ways in
which a user may be exposed to privacy threats.
In this paper, we expose and discuss a previously unconsidered
avenue by which a user’s privacy can be gravely exposed.
Using this exploit, we were able to gain complete access
to some popular online social network accounts without
using any conventional method like phishing, brute force, or
trojans. Our attack merely involves a legitimate exploitation
of the vulnerability created by the existence of obsolete webbased
email addresses. We present the results of an experimental
study on the spread that such an attack can reach,
and the ethical dilemmas we faced in the process. Last, we
outline our suggestions for defense mechanisms that can be
employed to enhance online security and thwart the kind of
attacks that we expose.
Categories and Subject Descriptors
K.4.0 [COMPUTERS AND SOCIETY]: General; K.4.1
[COMPUTERS AND SOCIETY]: Public Policy Issues—
Privacy
Keywords
Online social networking; Facebook; Phishing; Brute Force;
Identity; Media
Copyright is held by the International World Wide Web Conference
Committee (IW3C2). IW3C2 reserves the right to provide a hyperlink
to the author’s site if the Material is used in electronic media.
WWW 2013 Companion, May 13–17, 2013, Rio de Janeiro, Brazil.
ACM 978-1-4503-2038-2/13/05.
1. INTRODUCTION
Online social networks such as Orkut, Facebook, MySpace,
etc. have gained immense popularity over the recent
years. While facilitating communication and interaction
among their users, these networking platforms have also
raised increasing security and privacy concerns, as malicious
users, attackers, or hackers have been attempting to compromise
the confidentiality of users’ private information and
to gain access to other people’s accounts in an illegitimate
fashion.
For instance, LinkedIn, a popular social networking site
for professionals, recently came under attack by Russia-based
hackers who publicized the passwords of more than 6.5 million
users’ accounts [2]. Another related incident occurred in
February 2013 when Twitter, a popular micro blogging service,
was attacked, compromising the personal information
of more than 250,000 users [4] and exposing the vulnerability
of all its users.
As of December 2012, there are more than 1 billion monthly
active Facebook users [1], which roughly equals one-seventh
of the entire human population on our planet. Along with
the growth in the number of active users, which has been
following an exponential pattern, the size of personal data
stored on remote servers is also growing. The immense popularity
of such services arises from the fact that it offers
an convenient, easy, and reliable manner to maintain contact
with friends, relatives, and co-workers, and even reestablish
contact with long-lost former classmates, neighbors,
and other associates. People who sign up for these
services trust the system with their personal information.
While public awareness of privacy concerns and vulnerability
has been recently growing, many users remain incognizant
of the potential for their personal information to be used
or compromised by malicious attackers, and, in some cases,
the service providers themselves. Academic research has devoted
significant efforts in delineating the ways in which
users’ information can be shared, published, and used in
a privacy-preserving manner [7, 6] and to what extent an
attacker can exploit bogus accounts in order to gain information
[5, 8]. Nevertheless, there has not been a sufficient
investigation of the several ways by which malicious adversaries
may gain access to other people’s accounts.
Online social networking platforms such as Facebook have
vulnerabilities, which users should be protected against without
compromising the usability of the system. Due to such
vulnerabilities, there have been numerous hacking attempts
in the past on the website itself, and more such attempts are
expected to occur in the future; social networking services
constitute a target of malicious users and hackers who are
sometimes merely attracted by the mere existence of such
vulnerabilities itself. Motivated by this state of affairs, in
this paper we study the weaknesses of prevalent social networking
services and assess the extent to which they are
vulnerable to such online attacks. We decided to focus on
the security aspects of Facebook, due to the overwhelming
popularity of this particular platform. In the course of our
study, we eventually identified a security exploit, which, surprisingly,
allows an adversary to gain complete control over
a user’s Facebook account even without entering into hacking
activities per se. This identified threat is not limited
to Facebook only; the same concept can be applied to any
online web service which fulfills certain criteria.
Our exploit is not designed with the intention to target
any specific user. Instead, we search for, and exploit the
vulnerability of, users who used to possess a web-based email
account, which they used in order to sign up when creating
their personal account on Facebook in the first place, yet
those email accounts have in the meantime expired according
to the expiration rules of the web-based service that provided
them. This expiration is due to the fact that certain webbased
email providers configure the accounts they provide
to expire after a certain period of inactivity. Besides, some
users may decide to delete their own email accounts without
realizing the security threats that this action entails. Such
threats arise from the fact that the same web-based email
services allow any other willing user to reactivate and use the
same email address which had previously expired, when they
sign up. In our study, we found ourselves able to reactivate,
and thereby gain control of, such email address accounts;
thereafter, using the default password recovery mechanism
provided by Facebook, we were also able, in consequence, to
gain complete control over Facebook users’ private accounts.
In effect, the exploit we have identified carries the potential
to affect many users with complete loss of control over the
personal information.
2. THE ATTACK EXPERIMENT
We started out our study of the Facebook system’s security
using conventional hacking mechanisms like brute force.
We also tried certain social engineering methods such as
phishing, so as to see whether people may still fall into these
traps. Nevertheless, in the process, we realized the possibility
for a remarkably simple exploit which can give us access
to a user’s complete account and deny access to the same
account to that user herself. The potential victims of this
exploit are users who have originally created their Facebook
accounts using an email address which in the meantime expired
due to inactivity.
The exploit arises from the fact that, in order to set up
a Facebook account, users are required to provide an email
address. While some people opt to use their primary email
address to open up an account, others use their least used
or rarely used email address. In the case of the latter, the
email provider can apply a policy by which email accounts
expire after a period of inactivity; examples of such service
providers are web-based email services such as Hotmail; in
such cases, the user’s expired email, and, thereby, their Facebook
account as well, are up for grabs. In particular, once an
email account has expired due to inactivity, the inactivated
email address returns to the pool of available addresses; anyone
can then legitimately claim such an address when they
set up their own web-based email account. As a result, by
means of a very simple process of email account reactivation,
an email address that has previously belonged to another
person can be rendered ours.
The process we have outlined raises a question: How can
we detect email accounts that have expired. To facilitate
and automate this process, we developed a shell script which
checks theMX records on the mail server of any email provider
and sends a test email so as to check whether the email is
received or not. A failure to deliver the test mail suggests
that the email account does not exist on the mail server.
The only downside to this approach is that the email address
of an individual has to be known and tested manually
by the script. Several email providers, such as, in our case,
Hotmail, provide an even easier option to find not only one,
but a group of expired email accounts. Windows Live Messenger,
an instant messaging service provided by Microsoft,
allows anyone to import their friends list from Facebook.
The records in this imported list are categorized into two
groups:
1. People who are currently on Windows Live.
2. People who are not currently on Windows Live.
Membership in the first category signifies that the person
in question has already signed up for the Windows Live
service; besides, people having a Hotmail accounts are automatically
signed up for Windows live. On the other hand,
membership in the second category denotes that the person
in question does not currently hold an active Windows Live
account. Then, in case that person’s email is Hotmail email
address, we can safely conclude that this email address has
expired. We can then proceed to reactivate it ourselves.
Once we have acquired control of a previously expired
email address, which had once been used to open up a Facebook
account, we can visit Facebook on the web and claim
to be the user in question and have forgotten our password.
Facebook then promptly sends an email to our reactivated
Hotmail email address, which contains a code that allows us
to reset the password for the Facebook account in question.
All we need to do us copy the submitted code to a designated
field on the Facebook site. Once we have done so, we are
asked to set a new password. Then the Facebook account in
question is all ours, as we are now acting as the legitimate
owners of that account. Besides, this process can go on; we
can repeat it for every new account that can fall prey to our
attack method. Besides, by gaining access to more Facebook
accounts, we can automatize the process even further.
We can get the friends list of the account that we enter into
and figure out which of those friends have expired Hotmail
accounts of their turn. Thus, they fall prey to our attack as
well. This process can go on in a chain-reaction, branchingout
manner, accumulating more and more accounts that we
gain access to and deprive the original holders thereof from
access to them in the process. The process resembles building
a tree iteratively; at each iteration, the tree leaves are
the friends of users compromised in the previous iteration;
those leaves that can fall prey to our attack are “opened
up” and generate children-nodes in the next iteration. This
process would only encounter a dead-end when it reaches a
point where there are no more vulnerable leaf nodes. We
originally speculated that such a state of affairs might be
encountered in practice, as users using Facebook accounts
with an expired email address might be limited in number
and sparsely distributed. However, as we found out in our
experiment, such a state of affairs was never reached: We
could always detect new accounts that could be compromised
at each iteration. We only stopped when we decided
to do so on ethical grounds. We found this result to be quite
alarming.
3. RESULTS
We visualize the results of our attack experiment by a
tree; the internal nodes of the tree correspond to compromised
accounts that we have entered into, starting out from
an account of ours we originally had access to as the root; the
leaves correspond to accounts that were reached as friends
and were not compromised, either because they were not
vulnerable or because we decided not to pursue the exploit
further. We follow a depth-first approach in building the
tree, in order to illustrate the fact that out attack can proceed
unimpeded across multiple levels at several iterations.
Our experiment began with a user having around 760 friends
out of which 4 were susceptible to this exploit. In this manner,
we were able to gain access to a total of fifteen accounts
across six tree levels; the corresponding tree is visualized in
Figure 1. While we gained full access to the compromised accounts,
we did not manipulate any of their contents. Thereafter,
we decided to terminate our exploit as we had already
achieved our illustrative proof-of-concept purpose. Pursuing
the attack further would merely create problems to more
compromised users and raise ethical dilemmas and concerns
for us, not to mention potential legal problems. Still, the
last node in out attack had more than 2000 friends, 23 of
which were vulnerable to our attack. Thus, we saw a significant
potential for our attack to be carried along across
more iterations.
In Figure 1, aij |k denotes the node on the ith level of the
tree, j refers to the numbering of nodes on that level, and
k refers to the number of vulnerable children nodes which
are friends for the parent node. We further use the following
notations: NSU denotes a Non-Susceptible User, DNP
indicates a path that we Did Not Pursue any further, while
CN indicates the Current Node with 23 susceptible friends,
at which we decided to discontinue the attack.
Figure 1: Tree depicting compromised accounts
Overall, we found that up to 2% of a user’s friends were
generally susceptible to our exploit, with the average value
being close to 1%. Thus, for a user with 300 friends, the
chances are that 3 of those friends are vulnerable to our
exploit. Figure 2 shows the declared locations of the 15
users who accounts we compromised on a world map, using
drawing pins. Remarkably, just with a small set of 15
compromised accounts are attack was able to reach world
scale.
Figure 2: World map with users location
4. DEFENSE MECHANISM
Arguably, Facebook is not the only party to be blamed
for the possibility of this exploit. A big portion of the fault
lies within Hotmail and its policies. Hotmail is free to set its
own rules and policies regarding the expiration of its users’
email accounts after a certain period of inactivity. However,
such expiration should not lead to a privacy threat for the
people concerned by rendering a profile they have created
on a social networking website vulnerable to an attack. In
short, the problem arises from the fact that the privacy of a
user’s online social network account rests on the privacy of
one’s email account. Once the user loses the one, they can
lose the other as well.
Facebook can protect users from this exploit. The best
method, in our view, would be to eradicate the dependency
between Facebook and other service providers, in this case
email providers. It is true that resetting a password by
means of an activation code sent to the user’s email is an old
and widespread password resetting method. However, the
policies of certain email providers render this method problematic.
Facebook can easily generate its own self-contained
procedure for password reset that would not rely on thirdparty
dependencies. For example, a method similar to the
one used for determining who is tagged in an image could be
used. By this procedure, Facebook could present the users
with images of different friends they have and ask them to
name those present. Yet this method would have its own
limitations as some people have thousands of friends out of
which they might forget some. Another possibility would be
to use an SMS service in combination with the email address
procedure. Besides, like several other web-based services do,
there could be a security question that would be asked of
users who claim to have forgotten their passwords.
Last, as the information stored and shared on Facebook
is personal, users themselves should pay more attention to
which email addresses they use for identification purposes
when they create an account, and maintain those email accounts
carefully thereafter. In particular, a user should pay
special attention when using an email address provided by
an organization having a policy of email account expiration.
5. LIMITATIONS
While our exploit can potentially be quite dangerous, it
has its own limitations as well. By our method, an attacker
cannot target any specific user. As discussed earlier, only
certain users who are vulnerable to this attack can have
their accounts compromised. This limitation withholds the
choice of whom to pursue from the attacker. Besides, an attack
has to be initiated from the attacker’s friend list. The
attacker has to import her Facebook friend list in her Hotmail
account. Once imported, she can follow the leads and
repeat this process for the people who are vulnerable to this
attack. Thus, only Hotmail and Windows Live users are
currently susceptible to this type of attack. Once their Hotmail
account becomes inactive, it expires and allows others
to claim the email address. To our knowledge, no other popular
email account provider currently lets an account expire
if not accessed regularly.
The attack we have carried out raises legal and ethical
questions. As our intention was only to prove the potential
of this exploit rather than maliciously use other people‘s private
information, we stopped our pursuit once we attested
that we had accumulated sufficient evidence of its practicability.
Certainly, techniques such as IP spoofing, using
a proxy server, or using a public workstation could significantly
reduce the risk of tracing the attack back to its origin.
Yet our focus was on illustrating the process rather than taking
protective measures and launching a large-scale attack
as a hacker would do.
6. LEGAL AND ETHICAL ISSUES
In our exploit, we have been gaining access into accounts
and thereby to the friends lists therein. Those friends would
later become our next target nodes. Initially, we were thrilled
to find out how conveniently we could gain access to other
people’s accounts. We speculated following the footsteps
of Ron Bowes, an information security consultant who collected
and published the public data of 100 million Facebook
users in 2010. If we had done something similar, it would
have shown that very little privacy to talk about is afforded
to Facebook users.
Nevertheless, after some careful consideration of the ethical
dimensions involved, we decided to settle with only showcasing
the possibility of this attack in this paper. Therefore,
we stopped our exploration after successfully gaining access
to 15 accounts, which we thought sufficed to prove our point.
We neither collected nor published any of the personal data
we could access. Furthermore we did not change any other
recovery settings. Thus, the compromised users could regain
access to the account by using their cellphone number
or answering their security question. These settings were not
modified in any way or form. Indeed, we found out that, after
a few days, some of the exploited users had gained back
their accounts using these recovery mechanisms. We could
have gathered private data hiding behind multiple proxies
or secure sockets; we did not do so as we considered how we
would have felt if somebody had publicized our private lives
to a wide audience, and decided to follow the ethical maxim
that we should treat others as we would like to be treated
ourselves.
7. CONCLUSION
The growing popularity of Facebook has made it a common
target for hackers and attackers. Although such attempts
are usually hindered by the high security features of
the Facebook system, those that do make their way through
can pose a substantial threat to users’ online privacy. For research
purposes, we attempted to determine the possibility
of a quite simple exploit that requires no special hacker skills
and credentials. Our results have proven our speculations to
be true. We were able to gain total and unlimited control
of a user’s account merely relying on an expired email account.
The underlying reason for the potential of this attack
is Hotmail’s email account expiration policy in combination
with Facebook’s policy of allowing a password to by reset
by relying merely on a user’s given email address. Even
though Facebook should not by fully blamed for the possibility
of this attack, it could easily prevent it. All they have
to do is change their password resetting techniques at least
for users having a Hotmail email address. In other words,
the password of a Facebook user registered by a Hotmail
email address should only be reset by a combination of other
techniques, calling for the users to prove their identify via
proving the knowledge of the friends they have already connected
to, answering one or more security questions, or via
a combination of those techniques. Eventually, we conclude
that a majority of users who trust social networking websites
with their personal information have very little or no
control on how this information can be manipulated. Such
users need to be more aware of privacy and security threats,
as any potential leak may lead to grave consequences. Experience
has shown that malicious users who try to crack
other people’s accounts are quite persistent and usually do
end up compromising the users’ privacy [3].
8. REFERENCES
[1] 1 billion facebook users on earth: Are we there yet? Online at:
http://www.forbes.com/sites/limyunghui/2012/09/30/1-billionfacebook-
users-on-earth-are-we-there-yet/.
[2] 2012 LinkedIn hack. Online at:
http://en.wikipedia.org/wiki/2012 LinkedIn hack.
[3] Hackers attempting to crack 600,000 facebook accounts every
day. Online at: http://www.dailymail.co.uk/sciencetech/article-
2054994/Facebook-hackers-attempting-crack-600-000-accountsday.
html.
[4] Twitter says hackers may have compromised 250,000 accounts.
Online at:
http://www.forbes.com/sites/andygreenberg/2013/02/01/twittersays-
hackers-may-have-compromised-250000-accounts/.
[5] Lars Backstrom, Cynthia Dwork, and Jon Kleinberg. Wherefore
art thou r3579x?: anonymized social networks, hidden patterns,
and structural steganography. In WWW, 2007.
[6] Yi Song, Panagiotis Karras, Sadegh Nobari, Giorgos Cheliotis,
Mingqiang Xue, and St´ephane Bressan. Discretionary social
network data revelation with a user-centric utility guarantee. In
CIKM, 2012.
[7] Yi Song, Panagiotis Karras, Qian Xiao, and St´ephane Bressan.
Sensitive label privacy protection on social network data. In
SSDBM, 2012.
[8] Mingqiang Xue, Panagiotis Karras, Raissi Chedy, Panos Kalnis,
and Hung Keng Pung. Delineating social network data
anonymization via random edge perturbation. In CIKM, 2012.How to Hack into Facebook without being a Hacker
Tarun Parwani
Rutgers University
tarun.parwani@rutgers.edu
Ramin Kholoussi
Rutgers University
rk496@rutgers.edu
Panagiotis Karras
Rutgers University
karras@business.rutgers.edu
ABSTRACT
The proliferation of online social networking services has
aroused privacy concerns among the general public. The
focus of such concerns has typically revolved around providing
explicit privacy guarantees to users and letting users
take control of the privacy-threatening aspects of their online
behavior, so as to ensure that private personal information
and materials are not made available to other parties and
not used for unintended purposes without the user’s consent.
As such protective features are usually opt-in, users have to
explicitly opt-in for them in order to avoid compromising
their privacy. Besides, third-party applications may acquire
a user’s personal information, but only after they have been
granted consent by the user. If we also consider potential
network security attacks that intercept or misdirect a user’s
online communication, it would appear that the discussion
of user vulnerability has accurately delimited the ways in
which a user may be exposed to privacy threats.
In this paper, we expose and discuss a previously unconsidered
avenue by which a user’s privacy can be gravely exposed.
Using this exploit, we were able to gain complete access
to some popular online social network accounts without
using any conventional method like phishing, brute force, or
trojans. Our attack merely involves a legitimate exploitation
of the vulnerability created by the existence of obsolete webbased
email addresses. We present the results of an experimental
study on the spread that such an attack can reach,
and the ethical dilemmas we faced in the process. Last, we
outline our suggestions for defense mechanisms that can be
employed to enhance online security and thwart the kind of
attacks that we expose.
Categories and Subject Descriptors
K.4.0 [COMPUTERS AND SOCIETY]: General; K.4.1
[COMPUTERS AND SOCIETY]: Public Policy Issues—
Privacy
Keywords
Online social networking; Facebook; Phishing; Brute Force;
Identity; Media
Copyright is held by the International World Wide Web Conference
Committee (IW3C2). IW3C2 reserves the right to provide a hyperlink
to the author’s site if the Material is used in electronic media.
WWW 2013 Companion, May 13–17, 2013, Rio de Janeiro, Brazil.
ACM 978-1-4503-2038-2/13/05.
1. INTRODUCTION
Online social networks such as Orkut, Facebook, MySpace,
etc. have gained immense popularity over the recent
years. While facilitating communication and interaction
among their users, these networking platforms have also
raised increasing security and privacy concerns, as malicious
users, attackers, or hackers have been attempting to compromise
the confidentiality of users’ private information and
to gain access to other people’s accounts in an illegitimate
fashion.
For instance, LinkedIn, a popular social networking site
for professionals, recently came under attack by Russia-based
hackers who publicized the passwords of more than 6.5 million
users’ accounts [2]. Another related incident occurred in
February 2013 when Twitter, a popular micro blogging service,
was attacked, compromising the personal information
of more than 250,000 users [4] and exposing the vulnerability
of all its users.
As of December 2012, there are more than 1 billion monthly
active Facebook users [1], which roughly equals one-seventh
of the entire human population on our planet. Along with
the growth in the number of active users, which has been
following an exponential pattern, the size of personal data
stored on remote servers is also growing. The immense popularity
of such services arises from the fact that it offers
an convenient, easy, and reliable manner to maintain contact
with friends, relatives, and co-workers, and even reestablish
contact with long-lost former classmates, neighbors,
and other associates. People who sign up for these
services trust the system with their personal information.
While public awareness of privacy concerns and vulnerability
has been recently growing, many users remain incognizant
of the potential for their personal information to be used
or compromised by malicious attackers, and, in some cases,
the service providers themselves. Academic research has devoted
significant efforts in delineating the ways in which
users’ information can be shared, published, and used in
a privacy-preserving manner [7, 6] and to what extent an
attacker can exploit bogus accounts in order to gain information
[5, 8]. Nevertheless, there has not been a sufficient
investigation of the several ways by which malicious adversaries
may gain access to other people’s accounts.
Online social networking platforms such as Facebook have
vulnerabilities, which users should be protected against without
compromising the usability of the system. Due to such
vulnerabilities, there have been numerous hacking attempts
in the past on the website itself, and more such attempts are
expected to occur in the future; social networking services
constitute a target of malicious users and hackers who are
sometimes merely attracted by the mere existence of such
vulnerabilities itself. Motivated by this state of affairs, in
this paper we study the weaknesses of prevalent social networking
services and assess the extent to which they are
vulnerable to such online attacks. We decided to focus on
the security aspects of Facebook, due to the overwhelming
popularity of this particular platform. In the course of our
study, we eventually identified a security exploit, which, surprisingly,
allows an adversary to gain complete control over
a user’s Facebook account even without entering into hacking
activities per se. This identified threat is not limited
to Facebook only; the same concept can be applied to any
online web service which fulfills certain criteria.
Our exploit is not designed with the intention to target
any specific user. Instead, we search for, and exploit the
vulnerability of, users who used to possess a web-based email
account, which they used in order to sign up when creating
their personal account on Facebook in the first place, yet
those email accounts have in the meantime expired according
to the expiration rules of the web-based service that provided
them. This expiration is due to the fact that certain webbased
email providers configure the accounts they provide
to expire after a certain period of inactivity. Besides, some
users may decide to delete their own email accounts without
realizing the security threats that this action entails. Such
threats arise from the fact that the same web-based email
services allow any other willing user to reactivate and use the
same email address which had previously expired, when they
sign up. In our study, we found ourselves able to reactivate,
and thereby gain control of, such email address accounts;
thereafter, using the default password recovery mechanism
provided by Facebook, we were also able, in consequence, to
gain complete control over Facebook users’ private accounts.
In effect, the exploit we have identified carries the potential
to affect many users with complete loss of control over the
personal information.
2. THE ATTACK EXPERIMENT
We started out our study of the Facebook system’s security
using conventional hacking mechanisms like brute force.
We also tried certain social engineering methods such as
phishing, so as to see whether people may still fall into these
traps. Nevertheless, in the process, we realized the possibility
for a remarkably simple exploit which can give us access
to a user’s complete account and deny access to the same
account to that user herself. The potential victims of this
exploit are users who have originally created their Facebook
accounts using an email address which in the meantime expired
due to inactivity.
The exploit arises from the fact that, in order to set up
a Facebook account, users are required to provide an email
address. While some people opt to use their primary email
address to open up an account, others use their least used
or rarely used email address. In the case of the latter, the
email provider can apply a policy by which email accounts
expire after a period of inactivity; examples of such service
providers are web-based email services such as Hotmail; in
such cases, the user’s expired email, and, thereby, their Facebook
account as well, are up for grabs. In particular, once an
email account has expired due to inactivity, the inactivated
email address returns to the pool of available addresses; anyone
can then legitimately claim such an address when they
set up their own web-based email account. As a result, by
means of a very simple process of email account reactivation,
an email address that has previously belonged to another
person can be rendered ours.
The process we have outlined raises a question: How can
we detect email accounts that have expired. To facilitate
and automate this process, we developed a shell script which
checks theMX records on the mail server of any email provider
and sends a test email so as to check whether the email is
received or not. A failure to deliver the test mail suggests
that the email account does not exist on the mail server.
The only downside to this approach is that the email address
of an individual has to be known and tested manually
by the script. Several email providers, such as, in our case,
Hotmail, provide an even easier option to find not only one,
but a group of expired email accounts. Windows Live Messenger,
an instant messaging service provided by Microsoft,
allows anyone to import their friends list from Facebook.
The records in this imported list are categorized into two
groups:
1. People who are currently on Windows Live.
2. People who are not currently on Windows Live.
Membership in the first category signifies that the person
in question has already signed up for the Windows Live
service; besides, people having a Hotmail accounts are automatically
signed up for Windows live. On the other hand,
membership in the second category denotes that the person
in question does not currently hold an active Windows Live
account. Then, in case that person’s email is Hotmail email
address, we can safely conclude that this email address has
expired. We can then proceed to reactivate it ourselves.
Once we have acquired control of a previously expired
email address, which had once been used to open up a Facebook
account, we can visit Facebook on the web and claim
to be the user in question and have forgotten our password.
Facebook then promptly sends an email to our reactivated
Hotmail email address, which contains a code that allows us
to reset the password for the Facebook account in question.
All we need to do us copy the submitted code to a designated
field on the Facebook site. Once we have done so, we are
asked to set a new password. Then the Facebook account in
question is all ours, as we are now acting as the legitimate
owners of that account. Besides, this process can go on; we
can repeat it for every new account that can fall prey to our
attack method. Besides, by gaining access to more Facebook
accounts, we can automatize the process even further.
We can get the friends list of the account that we enter into
and figure out which of those friends have expired Hotmail
accounts of their turn. Thus, they fall prey to our attack as
well. This process can go on in a chain-reaction, branchingout
manner, accumulating more and more accounts that we
gain access to and deprive the original holders thereof from
access to them in the process. The process resembles building
a tree iteratively; at each iteration, the tree leaves are
the friends of users compromised in the previous iteration;
those leaves that can fall prey to our attack are “opened
up” and generate children-nodes in the next iteration. This
process would only encounter a dead-end when it reaches a
point where there are no more vulnerable leaf nodes. We
originally speculated that such a state of affairs might be
encountered in practice, as users using Facebook accounts
with an expired email address might be limited in number
and sparsely distributed. However, as we found out in our
experiment, such a state of affairs was never reached: We
could always detect new accounts that could be compromised
at each iteration. We only stopped when we decided
to do so on ethical grounds. We found this result to be quite
alarming.
3. RESULTS
We visualize the results of our attack experiment by a
tree; the internal nodes of the tree correspond to compromised
accounts that we have entered into, starting out from
an account of ours we originally had access to as the root; the
leaves correspond to accounts that were reached as friends
and were not compromised, either because they were not
vulnerable or because we decided not to pursue the exploit
further. We follow a depth-first approach in building the
tree, in order to illustrate the fact that out attack can proceed
unimpeded across multiple levels at several iterations.
Our experiment began with a user having around 760 friends
out of which 4 were susceptible to this exploit. In this manner,
we were able to gain access to a total of fifteen accounts
across six tree levels; the corresponding tree is visualized in
Figure 1. While we gained full access to the compromised accounts,
we did not manipulate any of their contents. Thereafter,
we decided to terminate our exploit as we had already
achieved our illustrative proof-of-concept purpose. Pursuing
the attack further would merely create problems to more
compromised users and raise ethical dilemmas and concerns
for us, not to mention potential legal problems. Still, the
last node in out attack had more than 2000 friends, 23 of
which were vulnerable to our attack. Thus, we saw a significant
potential for our attack to be carried along across
more iterations.
In Figure 1, aij |k denotes the node on the ith level of the
tree, j refers to the numbering of nodes on that level, and
k refers to the number of vulnerable children nodes which
are friends for the parent node. We further use the following
notations: NSU denotes a Non-Susceptible User, DNP
indicates a path that we Did Not Pursue any further, while
CN indicates the Current Node with 23 susceptible friends,
at which we decided to discontinue the attack.
Figure 1: Tree depicting compromised accounts
Overall, we found that up to 2% of a user’s friends were
generally susceptible to our exploit, with the average value
being close to 1%. Thus, for a user with 300 friends, the
chances are that 3 of those friends are vulnerable to our
exploit. Figure 2 shows the declared locations of the 15
users who accounts we compromised on a world map, using
drawing pins. Remarkably, just with a small set of 15
compromised accounts are attack was able to reach world
scale.
Figure 2: World map with users location
4. DEFENSE MECHANISM
Arguably, Facebook is not the only party to be blamed
for the possibility of this exploit. A big portion of the fault
lies within Hotmail and its policies. Hotmail is free to set its
own rules and policies regarding the expiration of its users’
email accounts after a certain period of inactivity. However,
such expiration should not lead to a privacy threat for the
people concerned by rendering a profile they have created
on a social networking website vulnerable to an attack. In
short, the problem arises from the fact that the privacy of a
user’s online social network account rests on the privacy of
one’s email account. Once the user loses the one, they can
lose the other as well.
Facebook can protect users from this exploit. The best
method, in our view, would be to eradicate the dependency
between Facebook and other service providers, in this case
email providers. It is true that resetting a password by
means of an activation code sent to the user’s email is an old
and widespread password resetting method. However, the
policies of certain email providers render this method problematic.
Facebook can easily generate its own self-contained
procedure for password reset that would not rely on thirdparty
dependencies. For example, a method similar to the
one used for determining who is tagged in an image could be
used. By this procedure, Facebook could present the users
with images of different friends they have and ask them to
name those present. Yet this method would have its own
limitations as some people have thousands of friends out of
which they might forget some. Another possibility would be
to use an SMS service in combination with the email address
procedure. Besides, like several other web-based services do,
there could be a security question that would be asked of
users who claim to have forgotten their passwords.
Last, as the information stored and shared on Facebook
is personal, users themselves should pay more attention to
which email addresses they use for identification purposes
when they create an account, and maintain those email accounts
carefully thereafter. In particular, a user should pay
special attention when using an email address provided by
an organization having a policy of email account expiration.
5. LIMITATIONS
While our exploit can potentially be quite dangerous, it
has its own limitations as well. By our method, an attacker
cannot target any specific user. As discussed earlier, only
certain users who are vulnerable to this attack can have
their accounts compromised. This limitation withholds the
choice of whom to pursue from the attacker. Besides, an attack
has to be initiated from the attacker’s friend list. The
attacker has to import her Facebook friend list in her Hotmail
account. Once imported, she can follow the leads and
repeat this process for the people who are vulnerable to this
attack. Thus, only Hotmail and Windows Live users are
currently susceptible to this type of attack. Once their Hotmail
account becomes inactive, it expires and allows others
to claim the email address. To our knowledge, no other popular
email account provider currently lets an account expire
if not accessed regularly.
The attack we have carried out raises legal and ethical
questions. As our intention was only to prove the potential
of this exploit rather than maliciously use other people‘s private
information, we stopped our pursuit once we attested
that we had accumulated sufficient evidence of its practicability.
Certainly, techniques such as IP spoofing, using
a proxy server, or using a public workstation could significantly
reduce the risk of tracing the attack back to its origin.
Yet our focus was on illustrating the process rather than taking
protective measures and launching a large-scale attack
as a hacker would do.
6. LEGAL AND ETHICAL ISSUES
In our exploit, we have been gaining access into accounts
and thereby to the friends lists therein. Those friends would
later become our next target nodes. Initially, we were thrilled
to find out how conveniently we could gain access to other
people’s accounts. We speculated following the footsteps
of Ron Bowes, an information security consultant who collected
and published the public data of 100 million Facebook
users in 2010. If we had done something similar, it would
have shown that very little privacy to talk about is afforded
to Facebook users.
Nevertheless, after some careful consideration of the ethical
dimensions involved, we decided to settle with only showcasing
the possibility of this attack in this paper. Therefore,
we stopped our exploration after successfully gaining access
to 15 accounts, which we thought sufficed to prove our point.
We neither collected nor published any of the personal data
we could access. Furthermore we did not change any other
recovery settings. Thus, the compromised users could regain
access to the account by using their cellphone number
or answering their security question. These settings were not
modified in any way or form. Indeed, we found out that, after
a few days, some of the exploited users had gained back
their accounts using these recovery mechanisms. We could
have gathered private data hiding behind multiple proxies
or secure sockets; we did not do so as we considered how we
would have felt if somebody had publicized our private lives
to a wide audience, and decided to follow the ethical maxim
that we should treat others as we would like to be treated
ourselves.
7. CONCLUSION
The growing popularity of Facebook has made it a common
target for hackers and attackers. Although such attempts
are usually hindered by the high security features of
the Facebook system, those that do make their way through
can pose a substantial threat to users’ online privacy. For research
purposes, we attempted to determine the possibility
of a quite simple exploit that requires no special hacker skills
and credentials. Our results have proven our speculations to
be true. We were able to gain total and unlimited control
of a user’s account merely relying on an expired email account.
The underlying reason for the potential of this attack
is Hotmail’s email account expiration policy in combination
with Facebook’s policy of allowing a password to by reset
by relying merely on a user’s given email address. Even
though Facebook should not by fully blamed for the possibility
of this attack, it could easily prevent it. All they have
to do is change their password resetting techniques at least
for users having a Hotmail email address. In other words,
the password of a Facebook user registered by a Hotmail
email address should only be reset by a combination of other
techniques, calling for the users to prove their identify via
proving the knowledge of the friends they have already connected
to, answering one or more security questions, or via
a combination of those techniques. Eventually, we conclude
that a majority of users who trust social networking websites
with their personal information have very little or no
control on how this information can be manipulated. Such
users need to be more aware of privacy and security threats,
as any potential leak may lead to grave consequences. Experience
has shown that malicious users who try to crack
other people’s accounts are quite persistent and usually do
end up compromising the users’ privacy [3].
8. REFERENCES
[1] 1 billion facebook users on earth: Are we there yet? Online at:
http://www.forbes.com/sites/limyunghui/2012/09/30/1-billionfacebook-
users-on-earth-are-we-there-yet/.
[2] 2012 LinkedIn hack. Online at:
http://en.wikipedia.org/wiki/2012 LinkedIn hack.
[3] Hackers attempting to crack 600,000 facebook accounts every
day. Online at: http://www.dailymail.co.uk/sciencetech/article-
2054994/Facebook-hackers-attempting-crack-600-000-accountsday.
html.
[4] Twitter says hackers may have compromised 250,000 accounts.
Online at:
http://www.forbes.com/sites/andygreenberg/2013/02/01/twittersays-
hackers-may-have-compromised-250000-accounts/.
[5] Lars Backstrom, Cynthia Dwork, and Jon Kleinberg. Wherefore
art thou r3579x?: anonymized social networks, hidden patterns,
and structural steganography. In WWW, 2007.
[6] Yi Song, Panagiotis Karras, Sadegh Nobari, Giorgos Cheliotis,
Mingqiang Xue, and St´ephane Bressan. Discretionary social
network data revelation with a user-centric utility guarantee. In
CIKM, 2012.
[7] Yi Song, Panagiotis Karras, Qian Xiao, and St´ephane Bressan.
Sensitive label privacy protection on social network data. In
SSDBM, 2012.
[8] Mingqiang Xue, Panagiotis Karras, Raissi Chedy, Panos Kalnis,
and Hung Keng Pung. Delineating social network data
anonymization via random edge perturbation. In CIKM, 2012.
0 comments:
Post a Comment
Type your Comment