|
The IBM Zurich Research Lab is the European branch of IBM's worldwide
Research division located in Rüschlikon, Zurich. We conduct
fundamental and applied research that contributes to IBM products,
services, and solutions. The Security
& Assurance group focuses on methods to ensure that
the benefits and convenience of networked computing continue
to outweigh the risks of operating in an open networked environment.
We are looking for talented students interested in working with
us on various projects, which can be conducted as a Master's thesis
project (6 months). Project starting date: mid-February/March 2008.
IBM is committed to diversity at the workplace. With us, you will
find an open, multicultural environment offering both women and
men excellent opportunities to develop their social and professional
skills. Students can explore an international research culture bringing
together academics and industry, understand how IBM is dedicated
to the success of its customers, learn about transferring research
ideas into products, and become familiar with the latest IT technologies
and novelties in IT services.
Interested students should apply directly by email or send inquiries
to the contact person given for each project.
| Project 1 |
Runtime monitoring of
timed software systems |
| |
Runtime monitoring of software systems is
concerned with the following idea: Given a correctness property
p, usually expressed in a propositional linear temporal
logic such as LTL, a runtime monitor is generated that recognizes
all state traces violating p. The system to be monitored
as well as the generated monitor are then executed in parallel.
State changes are communicated to the monitor as events such
that the behavior of the system can be observed and constrained.
Our focus lies on timed software systems and first-order
logic. We thus specify desirable system properties using
a subset of Metric First-Order Temporal Logic (MFOTL), a
highly-expressive language to describe static and dynamic
system properties. In spite of this expressivity, the runtime
monitoring problem for MFOTL is decidable [1]. This makes
runtime monitoring applicable to many interesting areas such
as regulatory compliance, usage control, business activity
monitoring, automated stock trading, or the checking of dynamic
integrity constraints for temporal databases.
With these
exciting applications in mind, we are looking for an motivated
student interested to implement, improve, and validate existing
MFOTL runtime monitoring approaches. The main tasks include
but are not necessarily limited to:
| • |
Implementation of the monitoring algorithm
given in [1]. This should be done either based
on standard technology such as Java and DB2 or, alternatively, based on a functional
programming
language. |
| • |
Experimental validation of complexity results
by investigating the performance of the decision algorithm
for typical formula classes. |
| • |
Improvement of the algorithm and contribution to open
theory problems (optional). |
| • |
Documentation of results. |
| |
|
| [1] |
D. Basin, F. Klaedtke, S. Müller, and B. Pfitzmann.
Runtime monitoring of metric first-order temporal logic
properties. Technical
Report RZ 3702, IBM Research and ETH Zurich, February 2008. |
|
| Prerequisites |
The optimal candidate for this topic has a well-balanced
interest in both systems and theory. In
particular, we expect:
| • |
Advanced practical programming skills. |
| • |
Knowledge of and interest in first-order logic. |
|
| Work environment |
The student selected for this project will work with a small
team of researchers at the IBM Zurich
Research Laboratory and at ETH. He or she will have the opportunity
to contribute to open research
problems, learn more about IBM products and tools, and work
in a friendly, multi-cultural
atmosphere. |
| Contact |
Andreas Wespi (anw@zurich.ibm.com) |
| |
 |
| Project 2 |
Multi-level secure service
broker |
| |
Multi-level security is based on a lattice
of security labels where secrecy (Bell/LaPadula) or integrity
(Biba) decreases from high to low. Basic flow control policies
then allow up-flow while disallowing down-flow. For example,
secret messages cannot be stored on public machines whereas
public messages can be sent to secret machines.
Although most MLS architectures are based on proprietary
software, they are likely to be migrated to service-oriented
architectures using Web services. A core component of a
SOA is a service broker that routes service requests to appropriate
service providers.
In order to allow this migration of multilevel secure systems
onto service-oriented architectures, the goal of this project
is to design and implement a prototype service broker that
enforces multilevel security policies.
The basic idea of our current design (see IBM CyberDigest,
RZ3672) is to run one broker per level while enforcing message
flow by placing gateway services between two security classifications/levels.
Challenges to overcome
are
| • |
service advertising across multiple levels, |
| • |
modifying existing service brokers to allow brokering
of multilevel services,
and |
| • |
synchronizing service requirements and
flow control policies. |
|
| Prerequisites |
Candidates for this project have a well-balanced
interest in both systems and theory. We expect creativity, good programming skills (e.g.,
Java, Eclipse), and the ability to document the work done. |
| Work environment |
The student selected for this project will work with a small
team of researchers at the IBM Zurich Research Laboratory.
He or she will have the opportunity to contribute to open research
problems, learn more about IBM products and tools, and work
in a friendly, multi-cultural atmosphere. |
| Contact |
Andreas Wespi (anw@zurich.ibm.com) |
| |
|
|
|
|
