Thesis Topics

On this page, you will find a list of available thesis topics that are available in our institute. Information about on-going and past theses can be found on this page. Some of the thesis descriptions are in German.

Note that because many of our topics are issued in German, some of the descriptions on this page are also German only. We are currently working on providing complete translations.

Open Theses

“Confidential Computing via Multiparty Computation and Trusted Computing,” Project or Master's thesis, D. Meißner (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
Secure multiparty computation and trusted computing are two very different approaches to perform computations on confidential data. MPC relies on provably secure but costly cryptography whereas trusted computing relies on somewhat trustworthy hardware. This difference also affects to the underlying attacker models. Simply put MPC requires non-colluding parties and trusted computer requires uncompromised hardware that was correctly manufactured. The goal of this topic is to explore a combination of both approaches to mitigate each others weaknesses for a specific use case where the computing parties are disjunct from data subjects.
“Extending ISOS by tentative decisions,” Master's thesis, F. J. Hauck (Supervisor), F. J. Hauck (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
ISOS is a leaderless consensus protocol for state-machine replication. It brings any input to the state machines in a total order so that it can be processed in a deterministic way. Typically requests to the machines are processed after the decision in the consensus protocol. As the consensus takes many message exchanges it needs some time. A first goal of this thesis is to implement a version of ISOS in our own framework called SMRteez. The protocol shall be extended by a to-be-defined interface that delivers preliminary decision to a processing unit earlier than the final decision. The interface needs to be able to deliver not only final decisions but also revokations of preliminary decisions, so that the execution engine can abort processing. The student has to identify the extension points when and where preliminary decisions can be derived. Further the new protocol is to be evaluated in order to verify how many wrong decisions are processed in which situations (high load, faulty nodes, etc.). The execution engine is outside of this work, it just needs to be simulated, e.g. by re-executing aborted requests on definite decisions.
“Graph Machine Learning on Evolving Graphs,” Master's thesis, B. Erb (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Event-sourced graphs are a concept to capture graphs that are continuously evolving, e.g. based on a stream of live changes. Goal of this thesis is (a) to provide an overview of existing graph machine learning concepts, (b) to assess the applicability of machine learning mechansims on event-sourced graphs, and (c) to prove the feasibility with a prototype implementation. Basic knowledge of general machine learning concepts is a plus for this thesis.
“Library Isolation in Embedded Systems,” Bachelor's thesis, Master's thesis, F. Kargl (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Software vulnerabilities pose critical threats to safety-relevant systems like vehicles. For example, a bug in a library embedded into a control program of a braking electronic-control-unit (ECU) inside a vehicle might be exploited by attackers to infiltrate the ECU and disable the breaks. More and more of today's vehicle allow over-the-air update of their software, but this typically requires updating a full firmware or at least a monolitic binary. So even if the mentioned library is patched, it still takes substantial time to integrate the new library version with the binary, and create and distribute a new firmware. In this work, we want to investigate system-level mechanisms to isolate a faulty library by dissecting it from a binary and executing it in a sandboxed environment. Feasibility of the concepts should be demonstrated in a proof-of-concept implementation.
“Test-Case-Generation Strategies for Network-Protocol-Model Refinements,” Bachelor's or Master's thesis, S. Kleber (Supervisor), F. Kargl (Examiner), Institut of Distributed Systems, Ulm University, 2019 – Open.
Security assessments of networked systems require knowledge about the utilized communication protocol. For proprietary protocols without known specification and with only limited access to the end-points, the only source of information is the communication itself. To correctly conclude from the captured byte stream to message-formats, -types, and finally a protocol model, structure, message- and field-boundaries, data-type, and semantics need to be inferred. After an initial inference procedure, it is desirable to refine the existing protocol model. Actively probing an entity for the validity of message syntaxes allows to targetedly enhance the knowledge about the protocol. To do this efficiently a smart method of automatically generating test-cases depending on the current protocol model needs to be developed.
Contact

Secretary's Office

Marion Köhler
E-Mail
Phone: +49 731 50-24140
available in the morning
Fax: +49 731 50-24142

Postal Address

Institute of Distributed Systems
Ulm University
Albert-Einstein-Allee 11
89081 Ulm

Visiting Address

James-Franck-Ring
Gebäude O27, Raum 349
89081 Ulm
manned in the morning Monday to Thursday

Directions

Topics By Degree

Bachelor's Theses

“Library Isolation in Embedded Systems,” Bachelor's thesis, Master's thesis, F. Kargl (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Software vulnerabilities pose critical threats to safety-relevant systems like vehicles. For example, a bug in a library embedded into a control program of a braking electronic-control-unit (ECU) inside a vehicle might be exploited by attackers to infiltrate the ECU and disable the breaks. More and more of today's vehicle allow over-the-air update of their software, but this typically requires updating a full firmware or at least a monolitic binary. So even if the mentioned library is patched, it still takes substantial time to integrate the new library version with the binary, and create and distribute a new firmware. In this work, we want to investigate system-level mechanisms to isolate a faulty library by dissecting it from a binary and executing it in a sandboxed environment. Feasibility of the concepts should be demonstrated in a proof-of-concept implementation.
“Test-Case-Generation Strategies for Network-Protocol-Model Refinements,” Bachelor's or Master's thesis, S. Kleber (Supervisor), F. Kargl (Examiner), Institut of Distributed Systems, Ulm University, 2019 – Open.
Security assessments of networked systems require knowledge about the utilized communication protocol. For proprietary protocols without known specification and with only limited access to the end-points, the only source of information is the communication itself. To correctly conclude from the captured byte stream to message-formats, -types, and finally a protocol model, structure, message- and field-boundaries, data-type, and semantics need to be inferred. After an initial inference procedure, it is desirable to refine the existing protocol model. Actively probing an entity for the validity of message syntaxes allows to targetedly enhance the knowledge about the protocol. To do this efficiently a smart method of automatically generating test-cases depending on the current protocol model needs to be developed.

Master's Theses

“Confidential Computing via Multiparty Computation and Trusted Computing,” Project or Master's thesis, D. Meißner (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
Secure multiparty computation and trusted computing are two very different approaches to perform computations on confidential data. MPC relies on provably secure but costly cryptography whereas trusted computing relies on somewhat trustworthy hardware. This difference also affects to the underlying attacker models. Simply put MPC requires non-colluding parties and trusted computer requires uncompromised hardware that was correctly manufactured. The goal of this topic is to explore a combination of both approaches to mitigate each others weaknesses for a specific use case where the computing parties are disjunct from data subjects.
“Extending ISOS by tentative decisions,” Master's thesis, F. J. Hauck (Supervisor), F. J. Hauck (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
ISOS is a leaderless consensus protocol for state-machine replication. It brings any input to the state machines in a total order so that it can be processed in a deterministic way. Typically requests to the machines are processed after the decision in the consensus protocol. As the consensus takes many message exchanges it needs some time. A first goal of this thesis is to implement a version of ISOS in our own framework called SMRteez. The protocol shall be extended by a to-be-defined interface that delivers preliminary decision to a processing unit earlier than the final decision. The interface needs to be able to deliver not only final decisions but also revokations of preliminary decisions, so that the execution engine can abort processing. The student has to identify the extension points when and where preliminary decisions can be derived. Further the new protocol is to be evaluated in order to verify how many wrong decisions are processed in which situations (high load, faulty nodes, etc.). The execution engine is outside of this work, it just needs to be simulated, e.g. by re-executing aborted requests on definite decisions.
“Graph Machine Learning on Evolving Graphs,” Master's thesis, B. Erb (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Event-sourced graphs are a concept to capture graphs that are continuously evolving, e.g. based on a stream of live changes. Goal of this thesis is (a) to provide an overview of existing graph machine learning concepts, (b) to assess the applicability of machine learning mechansims on event-sourced graphs, and (c) to prove the feasibility with a prototype implementation. Basic knowledge of general machine learning concepts is a plus for this thesis.
“Library Isolation in Embedded Systems,” Bachelor's thesis, Master's thesis, F. Kargl (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Software vulnerabilities pose critical threats to safety-relevant systems like vehicles. For example, a bug in a library embedded into a control program of a braking electronic-control-unit (ECU) inside a vehicle might be exploited by attackers to infiltrate the ECU and disable the breaks. More and more of today's vehicle allow over-the-air update of their software, but this typically requires updating a full firmware or at least a monolitic binary. So even if the mentioned library is patched, it still takes substantial time to integrate the new library version with the binary, and create and distribute a new firmware. In this work, we want to investigate system-level mechanisms to isolate a faulty library by dissecting it from a binary and executing it in a sandboxed environment. Feasibility of the concepts should be demonstrated in a proof-of-concept implementation.
“Test-Case-Generation Strategies for Network-Protocol-Model Refinements,” Bachelor's or Master's thesis, S. Kleber (Supervisor), F. Kargl (Examiner), Institut of Distributed Systems, Ulm University, 2019 – Open.
Security assessments of networked systems require knowledge about the utilized communication protocol. For proprietary protocols without known specification and with only limited access to the end-points, the only source of information is the communication itself. To correctly conclude from the captured byte stream to message-formats, -types, and finally a protocol model, structure, message- and field-boundaries, data-type, and semantics need to be inferred. After an initial inference procedure, it is desirable to refine the existing protocol model. Actively probing an entity for the validity of message syntaxes allows to targetedly enhance the knowledge about the protocol. To do this efficiently a smart method of automatically generating test-cases depending on the current protocol model needs to be developed.

Topics By Research Area

Networks

“Test-Case-Generation Strategies for Network-Protocol-Model Refinements,” Bachelor's or Master's thesis, S. Kleber (Supervisor), F. Kargl (Examiner), Institut of Distributed Systems, Ulm University, 2019 – Open.
Security assessments of networked systems require knowledge about the utilized communication protocol. For proprietary protocols without known specification and with only limited access to the end-points, the only source of information is the communication itself. To correctly conclude from the captured byte stream to message-formats, -types, and finally a protocol model, structure, message- and field-boundaries, data-type, and semantics need to be inferred. After an initial inference procedure, it is desirable to refine the existing protocol model. Actively probing an entity for the validity of message syntaxes allows to targetedly enhance the knowledge about the protocol. To do this efficiently a smart method of automatically generating test-cases depending on the current protocol model needs to be developed.

Mobile Systems

Distributed Computing

Privacy

“Confidential Computing via Multiparty Computation and Trusted Computing,” Project or Master's thesis, D. Meißner (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
Secure multiparty computation and trusted computing are two very different approaches to perform computations on confidential data. MPC relies on provably secure but costly cryptography whereas trusted computing relies on somewhat trustworthy hardware. This difference also affects to the underlying attacker models. Simply put MPC requires non-colluding parties and trusted computer requires uncompromised hardware that was correctly manufactured. The goal of this topic is to explore a combination of both approaches to mitigate each others weaknesses for a specific use case where the computing parties are disjunct from data subjects.
“Graph Machine Learning on Evolving Graphs,” Master's thesis, B. Erb (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Event-sourced graphs are a concept to capture graphs that are continuously evolving, e.g. based on a stream of live changes. Goal of this thesis is (a) to provide an overview of existing graph machine learning concepts, (b) to assess the applicability of machine learning mechansims on event-sourced graphs, and (c) to prove the feasibility with a prototype implementation. Basic knowledge of general machine learning concepts is a plus for this thesis.

IT Security

“Confidential Computing via Multiparty Computation and Trusted Computing,” Project or Master's thesis, D. Meißner (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
Secure multiparty computation and trusted computing are two very different approaches to perform computations on confidential data. MPC relies on provably secure but costly cryptography whereas trusted computing relies on somewhat trustworthy hardware. This difference also affects to the underlying attacker models. Simply put MPC requires non-colluding parties and trusted computer requires uncompromised hardware that was correctly manufactured. The goal of this topic is to explore a combination of both approaches to mitigate each others weaknesses for a specific use case where the computing parties are disjunct from data subjects.
“Library Isolation in Embedded Systems,” Bachelor's thesis, Master's thesis, F. Kargl (Supervisor), F. Kargl (Examiner), Inst. of Distr. Sys., Ulm Univ., 2021 – Open.
Software vulnerabilities pose critical threats to safety-relevant systems like vehicles. For example, a bug in a library embedded into a control program of a braking electronic-control-unit (ECU) inside a vehicle might be exploited by attackers to infiltrate the ECU and disable the breaks. More and more of today's vehicle allow over-the-air update of their software, but this typically requires updating a full firmware or at least a monolitic binary. So even if the mentioned library is patched, it still takes substantial time to integrate the new library version with the binary, and create and distribute a new firmware. In this work, we want to investigate system-level mechanisms to isolate a faulty library by dissecting it from a binary and executing it in a sandboxed environment. Feasibility of the concepts should be demonstrated in a proof-of-concept implementation.

Fault Tolerance

“Extending ISOS by tentative decisions,” Master's thesis, F. J. Hauck (Supervisor), F. J. Hauck (Examiner), Inst. of Distr. Sys., Ulm Univ., 2022 – Open.
ISOS is a leaderless consensus protocol for state-machine replication. It brings any input to the state machines in a total order so that it can be processed in a deterministic way. Typically requests to the machines are processed after the decision in the consensus protocol. As the consensus takes many message exchanges it needs some time. A first goal of this thesis is to implement a version of ISOS in our own framework called SMRteez. The protocol shall be extended by a to-be-defined interface that delivers preliminary decision to a processing unit earlier than the final decision. The interface needs to be able to deliver not only final decisions but also revokations of preliminary decisions, so that the execution engine can abort processing. The student has to identify the extension points when and where preliminary decisions can be derived. Further the new protocol is to be evaluated in order to verify how many wrong decisions are processed in which situations (high load, faulty nodes, etc.). The execution engine is outside of this work, it just needs to be simulated, e.g. by re-executing aborted requests on definite decisions.

Cloud Computing

Multimedia Communication

Miscellaneous Topics