Conference PaperPDF Available

Dynamic Location Information in Attribute-Based Encryption Schemes

Figures

Content may be subject to copyright.
Dynamic Location Information in
Attribute-based Encryption Schemes
Iwailo Denisow, Sebastian Zickau, Felix Beierle, and Axel K¨
upper
Service-centric Networking
Technische Universit¨
at Berlin |Telekom Innovation Laboratories |Berlin, Germany
iwailodenisow@mail.tu-berlin.de | {sebastian.zickau|beierle|axel.kuepper}@tu-berlin.de
Abstract—Attribute-based encryption (ABE) allows users to
encrypt (cloud) data with fine-grained Boolean access control
policies. To be able to decrypt the ciphertext, users need to have
a private key with the associated attributes. If the attributes
satisfy the formula, the plaintext can be recovered. In this paper,
ABE is extended with dynamic attributes. This allows attributes
to be added to an existing private key. A server component
named Attribute Authority is introduced. By using these dynamic
attributes, it is now possible to have the decryption depend
on data that changes often, such as location information of a
mobile device. Two schemes were developed that convert location
data into usable ABE attributes. To demonstrate our results, an
Android application was implemented and evaluated in a field
test.
Index Terms—Location-based Access Control, Attribute-based
Encryption Schemes, Dynamic Attribute Information, Cloud
Data.
I. INTRODUCTION AND MOTIVATI ON
There is a vast and growing amount of private and con-
fidential cloud data [1]. Often there is the need to share this
cloud data. Some of the newly emerging technologies simplify
the way this can be done. One relatively recent development
is attribute-based encryption (ABE) [2]. It allows the user to
specify an access policy within the process of data encryption.
The access policy can be used to control who has access to
the data. To determine if a user can access a file, the attributes
within user’s private key are checked against the access policy.
They can contain some of the characteristics of users, e.g., the
date of birth, gender or domain/application specific attributes,
such as the position in the company hierarchy, e.g., employee,
manager, or external contractor.
However, one of the main problems with ABE is that the
attributes that are given to a user are static. To add or change
an attribute, the user has to be given a new private key.
This prohibits the use of attributes that change often, such
as data related to time: time of day or days of employment.
It would also be very beneficial to be able to express that
certain files can only be decrypted in specified time periods.
Another property that changes often is the location of a mobile
user. Due to an ever-increasing amount of mobile devices,
users now have the ability to gather information about their
location at almost any time. This makes it desirable to use
this information as part of the encryption scheme, as a new
approach to location-based access control (LBAC) [3]. The
infrastructure of ABE currently has no facilities to support
location attributes. The main goal of this work is extend
the expressiveness of the access policies by implementing
dynamic location attributes. To support location attributes
in a cloud computing environment, where multiple users
need access to the attributes, a client-server architecture is
required. Since location information is primarily available on
mobile devices, a client application needs to be developed.
One common problem of location-based services (LBS) is
that the location information sent by a mobile device can
be manipulated. In this work, the reliability or verification of
positioning information are only of secondary importance, and
it is presumed that the locations are correct or that the involved
parts of the system can be trusted.
The following section will give an overview of the related
work and the basic technologies that have been used. Sections
III and IV describe our main contributions and the implemen-
tation. In Section V the solutions are evaluated. Section VI
concludes the paper and gives an outlook on future work.
II. RE LATE D WORK
This section will give an overview of related work and the
technologies used during the realization and evaluation.
A. Attribute-based Encryption
ABE is a family of asymmetric encryption schemes that
depend on the use of pairing-based cryptography. The first
scheme of its kind was devised by Sahai and Waters and
published in 2006 [4]. In an ABE scheme there are three
different types of keys. In a first step, a key authority gen-
erates a secret master key and associated public keys. The
key authority should be a trusted entity since it has access
to the secret master keys. These secret master keys can be
used to generate private keys. A private key can be used to
decrypt data. The public keys can be used to encrypt data,
and may be distributed freely. With ABE, it is possible to
encrypt data with a fine-grained access control policy. This
is achieved through the use of attribute sets that are given
to the users. Private keys, which can only be created with the
secret master key, contain such attribute sets. When encrypting
a file, the user needs to specify a policy, a Boolean formula
describing what attributes are required for decryption, e.g.,
((employee and hiringdate <2012) or manager). This policy
describes that a file is only accessible either by managers or
employees that have been hired before 2012. If the attribute set
of the private key satisfies this Boolean formula, then access to
the plaintext is granted. One big advantage of ABE is that the
generation of private keys and the decryption is decoupled.
Private keys generated with certain attributes will always
be able to decrypt data for which they fulfill the Boolean
formulas, without ever communicating to the key authority
again.
This is not always a desirable property, since that means
the revocation of keys is not possible without re-encrypting
the data or redistributing every private key. Newer ABE
approaches can account for this problem with revocation
through the use of a proxy [5] or through policy delegation [6],
which allows updating the policy of a file to a more restrictive
policy requiring only the public key. However, because of
the decoupled nature of ABE, the attributes are static. They
can only be changed by distributing a new key. This is why
attributes cannot easily be used for location information the
location of a mobile device changes frequently. An important
characteristic of ABE is the prevention of collusion attacks. As
an example, a file that is encrypted with the policy (Aand B)
and two users. The first user has a private key with the attribute
Awhile the second user has a private key with the attribute
B. Each one of them is not able to decrypt the file, since both
attributes taken separately don’t satisfy the formula. It is then
a desirable property of ABE that both users cannot combine
their keys to form a single key which will be able to decrypt
the file. And in fact, ABE does fulfill this property, called
collusion resistance, as described in [2]. There are two variants
of ABE. In Ciphertext-Policy ABE (CP-ABE), the Boolean
formula (also called the policy) is saved in the ciphertext (the
encrypted file). The attributes that need to satisfy are saved in
aprivate key. In Key-Policy ABE (KP-ABE), these two roles
are directly reversed: the private key holds the formula and the
ciphertext possesses attributes. There are five basic functions
in CP-ABE:
Setup - Generates a public key and an associated secret master key.
Encrypt - Encrypts given data with a policy. A public key is needed.
Keygen - Creates a private key with attributes. To do this a secret master
key is needed. This private key is associated with the secret master key
and its public keys.
Decrypt - Needs a ciphertext and a private key. Decrypts the ciphertext
only if the attributes in the private key satisfy the policy and the private
key is associated with the public key that has been used to initially
encrypt the file.
Delegate - Needs only a private key. Generates a new private key with
a subset of the attributes.
There are two types of atomic formulas: normal attributes
and numerical attributes. A normal attribute is just a string
of letters and digits that begins with a letter, e.g., employee,
manager, workgroup301, male, or female. A numerical at-
tribute however has a name and an associated value. As an
attribute in a private key it would look like this: hiringdate
= 2014,birthyear = 1990. Only integers in the range from 0
to 264 1are supported. In access policies these numerical
attributes can be compared to numbers with the following
operators: =, <, >, ,, e.g., hiringdate 2014,birthyear
= 2000,accesslevel 8. It is not possible to compare two
different numerical attributes of a user, so the following is not
possible: count children >count pets. The second argument
of a numerical comparison always needs to be a number that
is known at the time of encryption.
For the access policy itself, there are three possible oper-
ators: or,and, and of. These operators can be combined to
form complex formulas. A file might be encrypted with the
following access policy: ((employee and hiringdate <2014
and accesslevel >3) or (employee and accesslevel >5) or
manager). As is usual for asymmetric encryption algorithms,
ABE is supposed to be used as a key encapsulation scheme,
i.e., as a hybrid encryption scheme [7]. This is a typical
way to construct systems using asymmetric encryption since
the asymmetric part is usually very slow. This is especially
necessary for ABE, because of the computational complexity
and the resulting time needed to encrypt or decrypt these keys.
B. Dynamic Attributes
There have been a number of publications that discuss the
addition of dynamic attributes to ABE [8][9]. However, each
of these findings has some drawbacks.
For example, in Weber’s approach from 2009 [8], the
dynamic attributes are inserted into a specific point in the
policy tree. In a static subtree, only the static attributes are
saved, whereas in a dynamic subtree the dynamic attributes
are saved. Both subtrees need to evaluate to true, because
they are combined with an AND. These dynamic attributes are
actually part of the policy, but only this single construction is
made to be able to handle dynamic attributes. This lowers the
expressiveness of ABE by requiring that the dynamic attributes
need to evaluate positively or the file cannot be decrypted. An
example for a formula that would not be possible to implement
using this approach: (manager or (employee and (time >8 and
time <16))) where time is a dynamic attribute. In 2012, Weber
[10] proposed a way to combine location-based encryption
(LBE) [11] and ABE. This solution does not integrate the
dynamic information into the attributes or into the policies of
ABE. It rather acts as an additional layer on top of the normal
encryption and decryption. As explained earlier, ABE is used
in a hybrid encryption scheme and encrypts a relatively small
amount of data. This encrypted data is used as a symmetric key
for the encryption or decryption using symmetric algorithm,
such as AES. In the proposed addition of dynamic locations
to ABE by Weber this dynamic key is XORed with the result
of the LBE, resulting in a final key that is then used to encrypt
and decrypt. A problem with this solution is that all encrypted
files using this method then require users that want to decrypt
the file to be at the specified location. It is not possible to
make this dependent on a part of the formula. For example,
the policy manager or (employee and location:office) describes
that a file can only be decrypted if the person is a manager
or if the person is an employee that is currently in the office.
This policy cannot be represented with Weber’s approach.
C. ABE Implementations
There are multiple implementations for ABE. Since one of
the goals is to write a prototype application for a mobile An-
droid device, it is necessary to check if these implementations
can be run there. CP-ABE [2] was the first implementation of
an ABE scheme. It was written by John Bethencourt in 2007.
For its calculations the pairing-based cryptography library
[12] is used. While it only supports CP-ABE, it features a
parser to transform the Boolean formulas into an internal
format. This implementation also has support for numerical
attributes in the private key and their comparison operators in
the encryption policies. Besides supporting and and or CP-
ABE also supports the use of an of operator.
Libfenc [13] is a library for functional encryption schemes.
It has been written in 2010 and 2011 by Akinyele et al. and
supports both CP-ABE and KP-ABE. Since it reuses the code
of the parser from Bethencourt’s CP-ABE implementation, it
also support numerical attributes. This library is also written
in C/C++. Development on this library stopped in 2011.
With Charm [14], there is a follow-up project. The major
drawback for this library is, that, while there is parser for
policies, it does not support numerical attributes. Further,
the parser does not handle the of operator. The predicate-
based encryption library (PEBEL) [15] builds upon Charm. It
extends the functionality presented in Charm by implementing
a hybrid encryption scheme, where ABE is used as the key
encapsulation scheme. It also implements numerical attributes
and the accompanying operators.
The Java version of CP-ABE has been written in 2013 by
Junwei Wang [16]. Instead of the pairing-based cryptography
library it uses a Java port (JPBC [17]) of it. This port has
been completely written in Java, making this library a prime
candidate to use, since it can be deployed on any machine
running Java, including Android. The internal formats of
attributes and policies are described in Section III.
III. DESIGN
This section contains the design decisions that explain, how
dynamic attributes and location attributes in an ABE scheme
are used.
A. Encoding Location Information
One of our goals was to integrate location attributes into
ABE. To do this, the location has to be encoded in some way.
In the following, we list geocoding systems considered. They
encode the latitude and longitude of a position into another
format. Some of the beneficial properties of these formats will
be explained.
The most adequate solution for encoding location informa-
tion is Geohash [18]. In 2008, Gustavo Niemeyer published a
website1describing his work on encoding geolocations called
Geohash. It is used to transform GPS coordinates into a short
string. Being an open system its algorithm has been granting
into the public domain. As such it is one of the most used
systems for the encoding of locations. Geohashes are encoded
in base 32, with the alphabet consisting of twenty-six letters
from a-z and six digits from 2-7. Internal bits are used to
1http://geohash.org/
specify the locations. Each letter describes the state of 5 bits.
The Geohash of the location 52.51 latitude and 13.32 longitude
with 6-character precision is: u336xp. The Geohash itself does
not specify a point but an area, i.e., a square or a rectangle,
respectively.
How the encoding and decoding works is best explained
with an example. Let’s say we wanted to encode the location
that is at 52.51 latitude and 13.32 longitude. Because longitude
and latitude are encoded in the same way, only the calculation
for the latitude (52.51) will be shown in the example, see
Table I. For the first bit the range of the latitude is set to
the maximum (90, 90). Then depending on the latitude being
greater than the mid of both ranges, the first resulting bit is
set. The bit is only set when the latitude is greater than the
mid of both values and not set otherwise. The range for the
next iteration of the algorithm has to be adjusted. If the bit
has been set, the latitude has been greater than the mid and all
values below that are being discarded. If the bit has not been
set, the latitude is smaller and the mid and all values greater
are being discarded. This is done by replacing the value that
is to be discarded with the middle of the previous iteration.
TABLE I: Example of how to encode latitude into a Geohash
# Bit latitude min latitude mid latitude max bit result
1 -90 0 90 1
2 0 45 90 1
3 45 67.5 90 0
4 45 56.25 67.5 0
5 45 50.625 56.25 1
6 50.625 53.4375 56.25 0
... ... ... ... ...
The longitude and latitude part of the hashes are combined
by alternating between the two. The 0th and all other even
bits belong to the longitude, while the 1st and all other odd
bits belong to the latitude. This is done, because shortening the
resulting string or bit string still yields an area where the initial
location is in, but with a lower precision. Each bit added to the
bit string decreases the area that is inscribed into a Geohash
by half. Adding two bits, one to the longitude and one to the
latitude part, results in an area that is four times smaller than
before. When removing bits, the same is also true, removing
two bits increases the size of the area by the factor four.
This also means that areas that are in proximity to each other
geographically tend to have common prefixes of the resulting
Geohashes. The Geohash of the Ernst-Reuter-Platz in Berlin
is u336xp and Eberswalde in the north of Berlin is u33u9d.
While they are about 50km apart, they both share the common
prefix: u33. So at least the first 15 bits of their Geohash are the
same. Sometimes, locations close to each other do not share
a common prefix. This happens at the borders of a Geohash,
e.g., the meridian and the 180th meridian. However, since it is
relatively easy to calculate the neighboring Geohashes, these
can also be used to determine vicinity.
B. Dynamic ABE Attributes
Dynamic attributes are attributes that can be added to a
private key after it has been created. This was not intended to
be possible. The naive approach to adding dynamic attributes
is to regenerate the private keys every time an attribute needs
to be added. This would be relatively easy to implement,
but it would require that there is a database that contains
all the users and their information. This is because when a
private key is created again, the attributes that would need to
be created for the users need to be saved somewhere. This
approach would also require some kind of authentication
since one would need to know which user is requesting the
new private key. The following approach seeks to overcome
these constraints. Every private key has a unique value D
and a set of attributes. This unique value is the part of
the key that prevents collusion attacks. Using the value D
and the secret master key from which the private key was
generated it is possible to reverse some of the computations
and calculate the value used for the generation of attributes
(gr). This value makes it possible to generate additional
attributes. During the setup stage of Bethencourt’s ABE
scheme [2] a public key PUK is generated with the values:
PUK =G0, g, h =gβ, f =g1 , e(g , g)α. Where G0
is a bilinear group. The exponents αand βare random.
The master key MK is also generated and has the value:
MK = (β, gα). During the key generation stage of the
scheme a private key PRK gets assigned the value: PRK =
D=g(α+r),jS:Dj=gr·H(j)rj, D0
j=grj.
Where rand all rjare random and His a hash function.
Since rjis random for each attribute and gis known through
the public key, only the value gris needed to compute new
attributes. Based on the value Dof the private key, the
following rearrangement of the formula is possible:
D=g(α+r)(1)
Dβ=gα+r(2)
gr=Dβ/gα(3)
The resulting Equation (3) shows that only D,βand gα
are required to calculate gr. That means only the value D
of the private key and both values of the master key are
required to compute new attributes. This approach minimizes
the communication needed between the owner of the private
key and the key authority. If the key authority has the master
key only the Dvalue of the private key, and additional
data for the generation of the attributes, such as location
information, needs to be transferred. Generating a key with
hundreds or even thousands of attributes can take a lot of
time. With this approach only the added attributes need to
be computed, the already existing attributes in the private
key can be used without any alteration and are not touched
upon or modified. This system is designed to run without
any external authentication. The general assumption is that
generated attributes are considered to be publicly available.
Users can easily create a new private key with the generated
attributes itself. For example, a file is encrypted with the policy
currentTime <1451602800, which specifies that the file can
only be decrypted until the beginning of 2016. The attribute
currentTime = <current unix time>can be retrieved from the
key authority. Because currentTime is the only attribute that
is required to decrypt the file, anyone that has access to the
key authority can create a new private key with the generated
attributes from it. The file should thus be considered to be
publicly accessible. However, when the file is encrypted with
currentTime <1451602800 and employee and the attribute
employee cannot be generated by the key authority. In this way
it is ensured that users who want to decrypt the file already had
aprivate key beforehand. When encrypting a file that should
not be publicly accessible, additional restrictions in the policy
formula are needed to authorize the users who try to decrypt
the file.
C. Location Attributes
To understand how location information can be added as
an attribute, it is explained how the internal attributes of the
ABE implementation work.
1) Internal Format: The JCPABE library uses an internal
format that has been adopted from Bethencourt’s implemen-
tation [2]. There is only one the kof noperator. The operator
evaluates to true if at least kof the nsubformulas evaluate
to true. The atomic formula in a policy only evaluates to true
if the attribute with the same name is in the private key. The
internal format uses the reverse polish notation. The operations
that are described in Section II-A are implemented with this
operator. The and operator in this format is emulated by setting
the kto n, requiring that all subformulas evaluate to true. The
or operator is emulated by setting kto 1, thus requiring that
at least one of the subformulas evaluate to true.
2) Geohash Solution: To implement location attributes,
we utilize Geohashes [18]. As discussed in Section III-A,
they have multiple beneficial properties, such as a binary
representation, the length determines the accuracy, and they
share a common prefix.
These properties can be used in ABE. For the usage
of Geohashes in a policy the following syntax has
been chosen to specify a permitted location area:
<attributename>:<latitude>:<longitude>:<precision>.
The last part of this format specifies the precision of the
Geohash with the number of bits used. The process of
converting this string to the internal representation will be
explained with an example. A user wants to allow a file to
be decrypted in northern Poland and eastern Germany. He
needs to enter the latitude and longitude of a location and a
precision. In this case he choses Ernst-Reuter-Platz in Berlin,
Germany as a starting point (latitude: 52.51, longitude: 13.32)
and a precision of 10 bits. The resulting policy has the form:
location:52.51:13.32:10. With the latitude, the longitude
and the precision in bits a Geohash is created. The binary
representation of the Geohash has the form: 11010000
1 1.
For each bit in this sequence an internal attribute is created
in the form: <attributename>geohash <bitmarker>. The
bitmarker is just a string of 64 Xs (an arbitrarily defined
maximum precision for a Geohash) where the Xat the position
of the bit replaced with the value of the bit. The internal
representation of the given area is displayed in Listing 1 (to fit
the page, this example is given with 32 bits as the maximum
precision). These attributes are then combined with an and.
The internal representation the of operator is used.
Listing 1: Internal representation of a Geohash policy
location geohash 1xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash x1xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xx0xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxx1xxxxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxx0xxxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxxx0xxxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxxxx0xxxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxxxxx0xxxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxxxxxx1xxxxxxxxxxxxxxxxxxxxxxx
location geohash xxxxxxxxx1xxxxxxxxxxxxxxxxxxxxxx
10 o f 1 0
3) Geohashes as an Attribute: Since users also want to
be able to decrypt a file that has been encrypted with a
location attribute, the private key must contain Geohash at-
tributes. A location attribute during the key generation has the
following syntax: <attributename>:<latitude>:<longitude>.
The precision does not need to be specified, caused by the
property that longer Geohashes give a higher precision. The
user location is saved as precise as possible (64 bits). 64
bits would be enough to specify specific square centimeters.
The high precision has the consequence that each Geohash in
a private key uses 64 attributes. If a key is generated with
the attribute location:54.3665:18.6335 (Gdansk, Poland) then
the first 30 bits of the Geohash look like this: 1101000011
1100110011 0001000001. For every bit of the 64 bits one
attribute is created in the same way as in the policy. Note
that both Geohashes have a common prefix; the first 10 bits of
the private key location and the 10 bits of the Geohash in the
policy are the same. Since the location attribute in the private
key has a higher precision than the area, which is guaranteed
because the location attribute in a private key is always 64
bits and both Geohashes share a common prefix. The newly
created private key can be used to decrypt files that have been
encrypted with the example policy.
4) Problems: If users encrypt their files with a location,
then the assumption may be, that every private key that has a
location attribute of the vicinity may be able to decrypt the file.
This is not necessarily true. In Figure 1, there is a marker just
east of the central square. The marker represents the location
that was entered to create the Geohash. The marker is on the
edge of the Geohash, meaning that even locations just around
of this marker will not have the same Geohash. This happens,
because Geohashes operate on a fixed grid. It is not possible
to arbitrarily move the rectangle. A solution to this problem
could be to include all the neighboring Geohashes, as it is
done in the prototype.
D. Transformation Approach
Bethencourt’s implementation of CPABE [2] contains sup-
port for numerical attributes. With the help of these attributes
it is possible to add support for location attributes and corre-
sponding areas in a policy. The reason why these numerical at-
tributes can not be used directly, is that only the integers from
0to 264 1can be represented. To build upon this existing fea-
ture, it needs to be possible to transform the range of the lon-
gitude and the latitude to the range of the numerical attributes
with the condition that the comparison operators still works
as expected. A policy to check if a user is in a given rectangle
(latitudeMin, longitudeMin, latitudeMax, longitudeMax)can
then be defined the following way: (latitude latitudeMax
and latitude latitudeMin) and (longitude longitudeMax
and longitude longitudeMin). A possible transformation for
the longitude would be:
lon : [180,180] 0,264 1N
lon(x) = $x+ 180
360 ·264 1%
For the latitude the same function could be used. But since
the latitude only ranges from 90to 90the result can be of
a higher precision. Both transformations are surjective, since
every value from 0to 264 1can be the result. The functions
are not bijective, because the cardinalities of the domain and
the co-domain are not equal. That these transformations work
with the existing operations of the numerical attributes, it
needs to be shown that the following statements are true (apart
from rounding errors):
ablon(a)lon(b),for each ∗ ∈ {=, <, >, ,≤}
ablat(a)lat(b),for each ∗ ∈ {=, <, >, ,≤}
The major advantage of this approach would be that areas
are not limited to the grid of a Geohash and can be arbitrarily
defined. This would make the neighboring Geohash solution
obsolete, resulting in a reduced time needed for the encryption.
During the key generation two numerical attributes, one for
longitude and one for latitude, need to be created. Each nu-
merical attribute is represented by about 70 internal attributes.
This means for a location attribute in a private key about 140
internal attributes are needed, 76 more than the 64 attributes
in the Geohash solution. The Geohash key generation should
perform better.
IV. IMPLEMENTATION
During the realization, three different components have been
implemented: the JCPABE library, a server component named
Attribute Authority (AA), and a mobile client application
based on the Android platform.
A. JCPABE
Our library implements the five basic functions described
in Section II-A and the additional functionalities described
in Section III. For the implementation of this library it has
been decided to build upon the Java CPABE version. The
most compelling argument for Java CPABE is its portability.
Since every component has been written in pure Java 6, it
can easily be run on Android-OS. We call our implementation
JCPABE2. The following paragraphs will give an overview
2https://github.com/TU-Berlin-SNET/JCPABE
of the introduced additions and changes to the Java CPABE
library.
1) Policy Parser: As explained in Section III, the C im-
plementation that had been ported uses two different policy
representations. All Boolean formulas are transformed into
the internal format. Then this format, which is also human
readable, is used to encrypt the file. The Java port of the
CPABE implementation lacks this transformation. This is
why we re-implemented the parsing capabilities of the C
implementation in Java. It is now possible to traverse the parse
tree and generate the internal policy that is used by the ABE
part.
2) Attribute Parser: Adding more complex attributes such
as numbers and locations also requires the parsing of these
attributes when generating private keys. This is much simpler
than parsing the policy, since the attributes neither have
operations nor a certain order. Because of this simplicity, it
was possible to implement the parsing with the help of regular
expressions.
3) Hashing of Attributes: The origin implementation saved
the name of the attribute in plaintext. This is a problem,
because the name of an attribute may already disclose infor-
mation to an attacker. For example, an attribute in a policy
may be called ”london”. By analyzing the policy tree, it is now
very easy to determine that this attribute is a location attribute.
With plaintext attributes, it would also be possible to use the
internal representation of the location attribute to determine
which Geohash has been used to create these attributes in the
first place. With this information, an attacker may either go to
that location, or even try to spoof the location, to ultimately
access a file which he should not have access to. All names
of attributes in private keys and in the policy formulas of the
encrypted files are saved after they have been hashed.
B. Attribute Authority
The AA is an implementation of a key authority. It manages
the public keys and the secret master keys. It is called AA
because it is a web application that can distribute generated
attributes. There are two separate parts to the AA. There is a
REST API and a web application that serve as a Management
Tool (MT). Both use the same SQLlite database to retrieve
information.
1) REST API: The REST API is used for the communica-
tion with the client application. Its main functionality is to
distribute generated attributes. It also has an index for the
secret and public keys. In this system, there are two types of
users: authenticated and unauthenticated.Authenticated users
are associated with a username. An overview of the REST
API methods:
Unauthenticated users:
GET/connectiontest
GET/publickeys
GET/publickeys/<name of a key pair>
GET/publickeys/<name of a key pair>/attributes
POST/publickeys/<name of a key pair>/attributes
Authenticated users:
GET/authtest
GET/secretkeys
GET/secretkeys/<name of a key pair>
POST/secretkeys/<name of a key pair>
DELETE/secretkeys/<name of a key pair>
POST/privatekeys/<name of a key pair>
2) Management Tool: The MT is used to control everything
related to the RESTful service. It is not meant to be used by an
end user. The MT features user and key management. Users
as well as master keys and their associated public keys can be
created and deleted. It can be used to change passwords. The
most important feature for the attribute generation however is
the management of attributes for each key pair. This controls,
which attributes users can generate through the AAs REST
API. The name and the type of the attribute are needed for
creation. Currently there are four types of attributes possible:
static string, string, number, and location.
C. Android Application
A proof of concept Android application has been developed.
Since the application is tightly integrated into the architecture,
nearly every action communicates with an AA. There are five
entries in the main menu:
Attribute Authority - Lists all registered AAs. It is possible to add or
delete AAs. Adding an AA requires a name and a URL.
Master Keys - Displays owned secret master keys, stored on AA level.
Create and delete secret master keys.
Private Keys - Shows all locally stored private keys. Create (with add
ad-hoc or pre-defined attributes) and delete private keys.
Encrypt a File - Set a public key, chose a file, create a policy (including
area selection on a map); file is encrypted and stored locally.
Decrypt a File - Set a private key and a decrypted a file. Requests new
attributes from the AA.
V. EVALUATI ON
In this section, the initial goals are evaluated. The main
goal was to increase the expressiveness of ABE by adding
dynamic location attributes. A comparison between the two
implemented solutions is also given. To show that the devel-
oped system works, a field test was conducted. Results of the
performance tests are at the end of this section. Additional
results can be found in a previous article [19].
A. Expressiveness of ABE Policies
The expressiveness of the access control mechanism from
ABE is one of the subjects of this paper. But how does
ABE compare to other access-control mechanisms? In 1992,
Ferraiolo and Kuhn published a paper about role-based access
control (RBAC) [20]. In their paper, they describe an access
control mechanism where users are given roles, and objects
are only usable when having a certain role. ABE has greater
expressiveness than RBAC, because every RBAC system can
easily be expressed in ABE, i.e., every role is converted into
an attribute. Users that have roles are given a private key with
these attributes. Actions, files, and objects are secured with
ABE with policies of the following form: role1 or role2 or
role3 or ... The only thing that is more difficult to translate
is the concept of role hierarchies. Roles can be comprised of
other roles. The easiest solution is to add each sub-role of a
role as an attribute as well. Since ABE can express even more
with the help of the and and or operators, the access control
mechanism of ABE is superior to RBAC.
An advancement of RBAC is the eXtensible Access Control
Markup Language (XACML) [21]. It is an attribute-based
access control scheme, which means that it can permit or deny
the authorization based on arbitrary contextual information.
This can include: actions the user wants to perform, objects
the user wants to perform the action on, the identity or the
time or the location of a requester, and many more. With
XACML, it is possible to express many more scenarios than
can be expressed with ABE. The main differences include
the negation of attributes and the usage of functions on the
attributes. It is for example possible to express that a user
should not have a certain attribute at all. In ABE, it is not
possible to compare two numerical attributes, while this is
possible with XACML.
The scenarios that can be expressed in ABE are manifold.
ABE’s expressiveness can even outperform certain access
control models. However, in contrast to the compared systems
in this section, ABE has the advantage that the attributes are
verified on the client side of the application. Thus, some of
the processes of the ABE system can be done without any
communication.
B. Comparison of both Dynamic Location Attribute Ap-
proaches
Two different solutions for dynamic location attributes have
been implemented. One is a proposed solution by Weber [10],
the other has been detailed in this paper. The differences
between the two are best explained when evaluating both
solutions using an application scenario. In this scenario, the
electronic health records (EHR) of patients are saved in an
encrypted file. The encrypted file is only accessible by doctors
and nurses. The nurses additionally also have to be at the
hospital to access the file. For both approaches, the private
keys of both users are the same. The nurse has a nurse attribute
and the doctor has a doctor attribute. For our solution, the
ABE policy of the file would look like this: doctor or (nurse
and location:hospital) where location:hospital is an geographic
area that encompasses the hospital.
With Weber’s approach, the ABE policy would have the
following form: doctor or nurse. To decrypt a file with this
approach, a private key with satisfying attributes is not enough.
An LBE key is also required to recover the plaintext. For this,
the device used by the doctors and nurses also need to send
their location to a server. If the location is inside the area of
the hospital, then the right LBE key is returned.
The first major difference between these two approaches
is, that in Weber’s approach, the doctors also need to be at
the hospital to view the data. This is because the location
information is used to get an LBE key that is combined
with the symmetric key, requiring that every decryption also
specifies the LBE key. With the location attribute inside the
ABE policy this is not necessary. A doctor can access the file
without requiring him to send a location or even communicate
with an AA at all. However, in the case of a time critical
medical emergency, Weber’s approach can have an advantage.
A nurse wants to decrypt the EHR. The location of the nurse’s
device is sent to a server which then decides in fractions of
a second if the location is valid or not. The time needed for
the decryption is mostly dependent on the mobile device the
nurse uses. In this situation, our approach is inferior. When a
nurse tries to access the file during an emergency, the nurse’s
device will send the location to the AA to generate the missing
attributes.
And in the case of a location attribute, 64 internal attributes
need to be generated. In time critical situations, Weber’s
approach seems to be better suited. However, our approach
preserves the expressiveness of ABE, allowing location at-
tributes to be only used dependent on the policy formula.
C. Field Test
To prove that the developed system actually works, a field
test has been conducted. Both of the implemented approaches
have been tested. A file has been encrypted twice, once for
each system.
Fig. 1: Field testing positions / allowed area
For our approach, the file was encrypted with the policy: attr
and location:52.51304:13.32062:30. The file that was tested
with Weber’s approach has been encrypted with the policy
attribute. Additionally, an LBE key was entered during the
encryption. The allowed area for the LBE implementation
inside the AA has then been adjusted. For both approaches,
only one private key has been used. The private key had the
attribute attr. Both systems have been configured to only allow
the decryption in the same area. The area can be described
with the Geohash u336xp and is displayed in Figure 1. Also
pictured in the same screenshot are the locations of the tests.
At each location, an attempt to decrypt both files was made.
The test has been conducted on a Nexus 4 Android device.
Even being close to the edge of the allowed area still produced
the correct results. This is due to the usage of GPS, which is
the most accurate positioning mechanism the mobile phone
supports. A major difference between the two systems is the
time it takes to decrypt files.
With Weber’s approach, the ABE decryption is significantly
shorter, since only one attribute is needed to do that part of
the decryption. With the other solution, at least 31 attributes
are needed (one for the static attribute and at least 30 for the
location attribute). The response times by the server are also
much higher, since the AA needs to generate 64 attributes for
the private key.
D. Performance
To get an impression on how the number of location
attributes is influencing the time to decrypt a file a performance
test on a laptop with JCPABE was executed. The computer’s
attributes are: Intel CPU i5-2410M at 2.3GHz, 8GB RAM,
Java-Version 8u45. The test run was performed with 20to
210 location attributes. As Figure 2 is using a logarithmic
representation on the x-axis, the linear results are not obvious
at first sight. The different conjunctive location attributes were
used in a disjunction. An application scenario for such a large
amount of location attributes could be the usage of all medical
practices locations within a city, e.g., Berlin.
100101102103
0
2
4
6
8
Location Attributes
Decryption Time (in s)
Fig. 2: Decryption time with increasing policy length (or)
VI. CONCLUSION AND OU TL OO K
A system has been developed which users can use to
securely share private data. In particular, users can also specify
at which location the data is accessible. For this some addi-
tions to the existing ABE schemes have been designed. The
contributions are the dynamic attributes and the proposed way
to use location information integrated into an ABE scheme.
A client application for the Android platform enables users
to use these dynamic attributes for encryption and decryption
of mobile cloud files. An already existing approach has been
implemented and compared to the proposed system. During
the work a second solution arose on how location information
could be integrated into ABE. One problem with the way
the dynamic attributes work is that a malicious user may
accumulate attributes that are meant to be only used once.
One possible solution would be to let ABE run in a trusted
environment on the device. This environment could then make
sure that additional attributes are discarded after a single use.
ACK NOW LE DG ME NT
The work presented in this paper was performed in the context
of the Curcuma3project within the Software Campus4pro-
3http://curcuma-project.net/
4http://www.softwarecampus.de/, (F¨
orderkennzeichen 01IS12056), The au-
thors are responsible for the content of this paper.
gram. The project is funded by the German Federal Ministry
of Education and Research (BMBF)5.
REFERENCES
[1] S. Zickau, M. Slawik, D. Thatmann, S. Uhlig, I. Denisow, and
A. K¨
upper, “Tresor – towards the realization of a trusted cloud ecosys-
tem,” in Trusted Cloud Computing. Springer, 2014, pp. 141–157.
[2] J. Bethencourt, A. Sahai, and B. Waters, “Ciphertext-policy attribute-
based encryption,” in Security and Privacy, 2007. SP’07. IEEE Sympo-
sium on. IEEE, 2007, pp. 321–334.
[3] S. Zickau, D. Thatmann, T. Ermakova, J. Repschlager, R. Zarnekow,
and A. Kupper, “Enabling location-based policies in a healthcare cloud
computing environment,” in Cloud Networking (CloudNet), 2014 IEEE
3rd International Conference on. IEEE, 2014, pp. 333–338.
[4] V. Goyal, O. Pandey, A. Sahai, and B. Waters, “Attribute-Based Encryp-
tion for Fine-Grained Access Control of Encrypted Data,” in Proceedings
of the 13th ACM conference on Computer and communications security.
ACM, 2006, pp. 89–98.
[5] S. Jahid and N. Borisov, “Piratte: Proxy-based immediate revocation of
attribute-based encryption,” arXiv preprint arXiv:1208.4877, 2012.
[6] A. Sahai, H. Seyalioglu, and B. Waters, “Dynamic credentials and
ciphertext delegation for attribute-based encryption,” in Advances in
Cryptology–CRYPTO 2012. Springer, 2012, pp. 199–217.
[7] R. Cramer and V. Shoup, “Design and analysis of practical public-key
encryption schemes secure against adaptive chosen ciphertext attack,
SIAM Journal on Computing, vol. 33, no. 1, pp. 167–226, 2003.
[8] S. G. Weber, “Securing first response coordination with dynamic
attribute-based encryption,” in Privacy, Security, Trust and the Manage-
ment of e-Business, 2009. CONGRESS’09. World Congress on. IEEE,
2009, pp. 58–69.
[9] N. Doshi and D. Jinwala, “Updating attribute in cp-abe: A new ap-
proach,” IJCA Proceedings on International Conference in Distributed
Computing and Internet Technology 2013, vol. ICDCIT, pp. 23–28,
January 2013.
[10] S. G. Weber, “A hybrid attribute-based encryption technique supporting
expressive policies and dynamic attributes,Information Security Jour-
nal: A Global Perspective, vol. 21, no. 6, pp. 297–305, 2012.
[11] L. Scott and D. E. Denning, “A location based encryption technique and
some of its applications,” 2003.
[12] B. Lynn, “On the implementation of pairing-based cryptosystems,” Ph.D.
dissertation, Stanford University, 2007.
[13] M. Green, J. A. Akinyele, and M. Rushanan. libfenc: The Functional
Encryption library. [Online]. Available: http://code.google.com/p/libfenc
[14] J. A. Akinyele, C. Garman, I. Miers, M. W. Pagano, M. Rushanan,
M. Green, and A. D. Rubin, “Charm: A framework for rapidly proto-
typing cryptosystems,” Journal of Cryptographic Engineering, vol. 3,
no. 2, pp. 111–128, 2013.
[15] J. de Muijnck-Hughes. (2013) pyPebel. [Online]. Available: https:
//github.com/jfdm/pyPEBEL
[16] J. Wang. (2013) CPABE. [Online]. Available: https://github.com/
junwei-wang/cpabe
[17] A. De Caro and V. Iovino, “jPBC: Java pairing based cryptography,”
in Proceedings of the 16th IEEE Symposium on Computers and
Communications, ISCC 2011. IEEE, 2011, pp. 850–855. [Online].
Available: http://gas.dia.unisa.it/projects/jpbc/
[18] G. Niemeyer. (2008) Geohash. [Online]. Available: http://geohash.org
[19] S. Zickau, F. Beierle, and I. Denisow, “Securing mobile cloud data with
personalized attribute-based meta information,” in Proceedings of the 3rd
IEEE International Conference on Mobile Cloud Computing, Services,
and Engineering. IEEE, 2015, pp. 205–210.
[20] D. Ferraiolo and R. Kuhn, “Role-based access controls,” Proc. of 15th
NIST-NSA National Computer Security Conference, 1992.
[21] OASIS Standard. (2013) eXtensible Access Control Markup Language
(XACML) Version 3.0. [Online]. Available: http://docs.oasis-open.org/
xacml/3.0/xacml-3.0- core-spec- os-en.html
5http://www.bmbf.de
... 2. Dynamic ABE Access Control Methods: The dynamic attributes of a mobile device are incorporated in the access control methods to overcome the above-discussed limitations. The dynamic location-based ABE access control approaches are proposed in [17][18][19]. Location of the mobile device is used by the data owners to encrypt their data. The approaches described in [17,18] follow SA-ABE in which the users' privacy may breach when the AA gets compromised. ...
... Location of the mobile device is used by the data owners to encrypt their data. The approaches described in [17,18] follow SA-ABE in which the users' privacy may breach when the AA gets compromised. This limitation has been overcome in [19] by incorporating the MA-ABE. ...
... Type of ABE No. of DAs Network disconnection Anonymity Location-based [17] SA-ABE One No No Location-based [18] SA-ABE One No No Location-based [19] MA-ABE One No Yes Lightweight access control [20] MA-ABE ...
... 2. Dynamic ABE Access Control Methods: The dynamic attributes of a mobile device are incorporated in the access control methods to overcome the above-discussed limitations. The dynamic location-based ABE access control approaches are proposed in [17][18][19]. Location of the mobile device is used by the data owners to encrypt their data. The approaches described in [17,18] follow SA-ABE in which the users' privacy may breach when the AA gets compromised. ...
... Location of the mobile device is used by the data owners to encrypt their data. The approaches described in [17,18] follow SA-ABE in which the users' privacy may breach when the AA gets compromised. This limitation has been overcome in [19] by incorporating the MA-ABE. ...
... Type of ABE No. of DAs Network disconnection Anonymity Location-based [17] SA-ABE One No No Location-based [18] SA-ABE One No No Location-based [19] MA-ABE One No Yes Lightweight access control [20] MA-ABE ...
Article
To strengthen the security of access control protocols for mobile cloud environment, dynamic attributes of mobile devices are used. The weak or disconnection issue of the mobile network is a critical task to deal with. The proposed approach provides access control as well as data confidentiality using dynamic attributes encryption. The pairs of mobile agents are used to deal with the issue of network connection. The secret key is distributed using the anonymous key-issuing protocol which preserves the anonymity of the user. The approach is implemented in a real mobile cloud environment, and the performance under various parameters is evaluated.
... Within the scope of this article use cases are addressed, in which private cloud data and IOT messages are secured and shared among friends in social and professional contexts or between sensors and requesting services. Additionally the expressiveness of ABE is expanded to also include dynamic attributes such as location information of mobile devices and time constraints [3]. ...
... There are use cases within the healthcare data domain that are based on ABE ciphertext-policies to protect electronic health records (EHR), namely Wang et al. [11], Alshehri et al. [12] and Akinyele et al. [13]. Within the project Curcuma 1 ABE is deployed in a location-based access control scenario whereas the location information is encoded inside the ABE attributes, see Zickau et al. [14] and Denisow et al. [3]. There exist solutions where ABE is applied within publish subscribe messaging-patterns (pub/sub-MP) [2]. ...
... The library has been tested to run on Android devices. The additions to the functionality are described in [3]. They include the generation and appending of attributes to an existing private key as well as an experimental feature that enables dynamic location information within policies. ...
... For resolving this issue, MA-CP-ABE schemes have been proposed where multiple AAs participate in the key-generation process [7]. Each AA is responsible for verifying a disjoint set of attributes and providing partial keys to users based on the users' provided attributes [8]. These individual AAs are independent, and there is no coordination among them during the key generation process. ...
... The use of dynamic attributes such as location, and unlock failures has been realized in [29], [30], [31], [32]. Strategies presented in [8], [29] uses only location of the user to encrypt and decrypt ciphertext. However, there is only a single authority, resulting in a user's privacy breach when the authority is compromised. ...
Article
Full-text available
Sensitive Data requires encryption before uploading to a public cloud. Access control based on Attribute-Based Encryption (ABE) is an effective technique to ensure data shared security and privacy in the public cloud. Cipher-Text Policy of ABE may suffer from scalability and performance issues as they do not permit for addition or removal of computing nodes at run time. Furthermore, another problem is existing approaches suffer from single-point-of-failure (SPoF). Therefore, we introduce a scalable multi-agent system architecture based on CP-ABE to ensure data sharing on public cloud storage and reliability in our proposed work. We proposed a cloud host as an inter-mediator between the user and the authorized agents without violating the system’s privacy and security. We have also proposed a novel methodology to protect the cloud from malware by exploiting the state’s efficient power of the art Gemini approach. Gemini is an efficient methodology for binary code-based graph embedding similarity detection. Our proposed study overcomes the deficiencies of scalability and efficiency along with providing the mechanism for malware detection in the cloud. It covers three aspects: scalability, the efficiency with multi-agents, and malware detection capability. Our contributed work is scalable, efficient for cloud data sharing, and protects from malware. Results reveal that our work provides better performance with preserving the security, privacy, and fine granularity features of CP-ABE and malware screening using regress analysis by the graph embedding technique.
... Given a simulator A, if T = e(g, g) abc , then CT * is a valid ciphertext, Adv = ε and P r A g, g a , g b , g c , T = e(g, g) abc = 0 = 1 2 + ε. (14) If T = R thenC is nothing more than a random value to the adversary. Therefore, ...
Article
Full-text available
Cloud computing has changed the way enterprises store, access and share data. Data is constantly being uploaded to the cloud and shared within an organization built on a hierarchy of many different individuals that are given certain data access privileges. With more data storage needs turning over to the cloud, finding a secure and efficient data access structure has become a major research issue. With different access privileges, individuals with more privileges (at higher levels of the hierarchy) are granted access to more sensitive data than those with fewer privileges (at lower levels of the hierarchy). In this paper, a Privilege-based Multilevel Organizational Data-sharing scheme~(P-MOD) is proposed that incorporates a privilege-based access structure into an attribute-based encryption mechanism to handle these concerns. Each level of the privilege-based access structure is affiliated with an access policy that is uniquely defined by specific attributes. Data is then encrypted under each access policy at every level to grant access to specific data users based on their data access privileges. An individual ranked at a certain level can decrypt the ciphertext (at that specific level) if and only if that individual owns a correct set of attributes that can satisfy the access policy of that level. The user may also decrypt the ciphertexts at the lower levels with respect to the user's level. Security analysis shows that P-MOD is secure against adaptively chosen plaintext attack assuming the DBDH assumption holds.The comprehensive performance analysis demonstrates that P-MOD is more efficient in computational complexity and storage space than the existing schemes in secure data sharing within an organization.
Conference Paper
Full-text available
Cloud computing is the flexible platform to outsource the data from local server to commercial cloud. However cloud provides tremendous benefits to user, data privacy and data leakage reduce the attention of cloud. For protecting data privacy and reduce data leakage various techniques has to be implemented in cloud. There are various types of cloud environment, but we concentrate on Hybrid cloud. Hybrid cloud is nothing but combination of more than two or more cloud. Where critical operations are performed in private cloud and non critical operations are performed in public cloud. So, it has numerous advantages and criticality too. In this paper, we focus on data security through encryption scheme over Hybrid Cloud. There are various encryption schemes are close to us but it also have data security issues. To overcome these issues, Attribute Based Encryption Scheme with Dynamic Attributes Supporting (ABE-DAS) has proposed. Attribute based Encryption Scheme with Dynamic Attributes Supporting technique enhance the security of the data in hybrid cloud.
Article
Full-text available
In mobile cloud computing (MCC), data are stored and processed by more powerful computing platforms located in clouds to enhance the capability of mobile devices. Although MCC brings many benefits to users and organizations, security has become the main concern for MCC because it involves third party environment where data confidentiality is not a guarantee. To overcome this challenge, attribute-based encryption (ABE) technique, a promising cryptography method was introduced to support both data confidentiality and access control simultaneously. However, most of the ABE used single certification authority (CA) and disjoint attribute authority (AA) to verify user identity which can be a single point of failure. Therefore, a new approach was proposed in this research to deal with the single point of failure in the authority. Furthermore, the cache-based scheduling algorithm was upgraded to improve the response time of user jobs. The experimental results show that the approach performed better compared to its competitor in terms of reliability. Additionally, it significantly reduced the number of read operations compared to its counterpart.
Conference Paper
Full-text available
In a multi-stakeholder cloud computing environment, data access control is of essential importance. Nowadays, it is usually handled in and deployed by every single cloud service on its own which makes the configuration of fine-grained access privileges cumbersome and economically expensive. In this paper, we introduce a novel cloud ecosystem architecture featuring an overall lightweight data access control model. This model is enabling data access policies based on location information of service consumer devices. We apply our architecture in the sensitive healthcare domain, which itself comprises multiple parties with complex data access privileges. Here, we define high-level requirements driven from current data protection regulations and guidelines as well as practice requirements in this area, which we address in the design of our architecture. We implement and test the main components. The results demonstrate the feasibility of our architecture and the applicability of our approach even in the healthcare application domain.
Article
Full-text available
BIOGRAPHY Logan Scott is a consultant specializing in radio frequency signal processing and waveform design for communications, navigation, radar, and emitter location. He has more than 24 years of military GPS systems engineering experience. As a senior member of the technical staff at Texas Instruments, he pioneered approaches for building high-performance, jamming-resistant digital receivers using large-scale application-specific integrated circuit (ASIC) technologies. He has developed gain and frequency plans, non-uniform analog/digital conversion techniques, fast acquisition architectures, Baseband signal processing algorithms and adaptive array approaches. He is currently involved in projects to provide location based encryption and authentication. He holds 27 US patents. ABSTRACT Location based encryption enhances security by integrating position and time into encryption and decryption processes. We find that from a security perspective, it is not enough to simply enable or disable decryption based on location and time; these aspects must be integrated into the key construction process. Furthermore, keys or files in transit should not reveal anything regarding their locations/times of applicability. After reviewing the objectives of location-based encryption, this paper introduces a specific approach called geo-encryption. The described geo-encryption approach builds on established cryptographic algorithms and protocols in a way that provides an additional layer of security beyond that provided by conventional cryptography. It allows data to be encrypted for a specific location(s) or for specific area(s), e.g. a corporation's campus area. Constraints in time as well as location can also be enforced. Geo-encryption can be used with both fixed and mobile applications and supports a wide range of data sharing and distribution policies. We then discuss a process of applying successive geo-encryptions at the originating node to enforce specific geographic routings for transmission to the final destination node. With each intervening node removing one layer of encryption, unless the file has gone through the proper sequence of nodes, decryption will fail. Using a similar process, messages can be location authenticated by applying one layer of encryption at each intervening node. Next, we discuss some specific applications. In the civilian sector, there has been a great deal of interest in providing location-based security for digital cinema distribution and forensic analysis in cases of piracy. In this application, the same, large (25 to 190 Gbyte), encrypted media file might be used at multiple theatre locations but with distinct GeoLocked keys specific to the ION NTM 2003, 22-24 January 2003, Anaheim, CA 734 intended recipient location and exhibition license. This provides a secure and efficient point to multipoint distribution model applicable to distributions via satellite or DVD. At the exhibition hall, robust watermarking/steganographic techniques can introduce location, time and exhibition license information into the exhibition for subsequent use in piracy investigations. For the military GPS user, we show how individual waypoints can be uniquely encrypted so as to be accessible only when the set is physically within the route parameters; both location and time wise. An intact, captured set would not reveal mission parameters.
Article
Full-text available
This article proposes a novel hybrid attribute-based encryption technique. It is attribute-based, as it allows encrypting under logical combinations of attributes, that is, properties that users satisfy. It is hybrid, as it combines ciphertext-policy attribute-based encryption (CP-ABE) with location-based encryption (LBE) on the level of symmetric keys. It supports encryption under expressive policies, since it can efficiently handle dynamic attributes with continuous values, such as location.
Article
Full-text available
In Ciphertext-Policy Attribute Based Encryption (CP-ABE), attributes are attached to the user's secret key and access policy is at-tached to the ciphertext. If attributes in the secret key of a user satisfy the policy then only the genuine user can decrypt the ciphertext. However, such scenario also necessitates periodic updating of the secret key with the changing attributes. According to our observations, the existing attempts at doing so are not efficient. In this paper, we propose a newer approach to add, update or delete the value of particular attribute effi-ciently without the knowledge of the other attributes.
Conference Paper
The TRESOR (Trusted Ecosystem for Standardized and Open cloud-based Resources http:// www. cloud-tresor. de/ ) project enables cloud computing solutions for the German health sector. This sector deals with sensitive medical information and is in general not suitable for current cloud-based solutions, which are lacking appropriate privacy and security features. The project evaluates and proposes new architectural components to address these shortcomings. These will be combined into a secure and trustworthy ecosystem that will enable the health industry and other sectors to take advantage of cloud computing. The architecture consists of components, such as a marketplace, a broker, a proxy and a PaaS-platform. TRESOR addresses privacy and data protection issues and aims at providing a standardized solution with reduced lock-in effects that can also be used in other domains. In this paper the specific tasks and the architecture of these components are presented, important challenges of the TRESOR project are highlighted and preliminary results, such as a secure transfer protocol, and policy integration are shown.
Article
Motivated by the question of access control in cloud storage, we consider the problem using Attribute-Based Encryption (ABE) in a setting where users’ credentials may change and ciphertexts may be stored by a third party. Our main result is obtained by pairing two contributions: We first ask how a third party who is not trusted with secret key information can process a ciphertext to disqualify revoked users from decrypting data encrypted in the past. Our core tool is a new procedure called ciphertext delegation that allows a ciphertext to be ‘re-encrypted’ to a more restrictive policy using only public information. Second, we study the problem of revocable attribute-based encryption. We provide the first fully secure construction by modifying an attribute-based encryption scheme due to Lewko et al. [9] and prove security in the standard model. We then combine these two results for a new approach for revocation on stored data. Our scheme allows a storage server to update stored ciphertexts to disqualify revoked users from accessing data that was encrypted before the user’s access was revoked while key update broadcasts can dynamically revoke selected users.
Article
We describe Charm, an extensible framework for rapidly prototyping cryptographic systems. Charm provides a number of features that explicitly support the development of new protocols, including support for modular composition of cryptographic building blocks, infrastructure for developing interactive protocols, and an extensive library of re-usable code. Our framework also provides a series of specialized tools that enable different cryptosystems to interoperate. We implemented over 40 cryptographic schemes using Charm, including some new ones that, to our knowledge, have never been built in practice. This paper describes our modular architecture, which includes a built-in benchmarking module to compare the performance of Charm primitives to existing C implementations. We show that in many cases our techniques result in an order of magnitude decrease in code size, while inducing an acceptable performance impact. Lastly, the Charm framework is freely available to the research community and to date, we have developed a large, active user base.