TUM Logo

Verifiable Computation

Verifiable Computation  

Seminare 2 SWS / 5 ECTS
Veranstalter: Claudia Eckert
Zeit und Ort:

Preliminary Meeting: Thursday, 13 July 2023 / 11.00 h / Room 01.08.033

Slides for preliminary meeting


Betreuer:innen Seminar Fraunhofer AISEC:

Johannes Wiesböck

Barbora Hrda

Katharina Bogad

Email: vorname.nachname@aisec.fraunhofer.de


The lecture is given in english

Advancements in network connectivity of edge devices motivate offloading
computations from the edge to cloud computing platforms for their higher and
cheaper computing power. Vice-versa, in cyber physical systems computations may
be offloaded to edge-devices for their proximity to a facility. In both cases,
machines ultimately executing computations may not be entirely under control of
the user which makes computation results susceptible to errors and manipulation.

This problem motivates for countermeasures in order to protect the integrity of
computations. Common solutions ensuring integrity include examining samples or
executing computations multiple times and comparing their results (e.g. BOINC).
State-of-the-art solutions rely on a setup consisting of trusted hardware and
software. Remote attestation features based on the Trusted Platform Module (TPM),
TCG DICE or Confidential Computing assure that a remote setup is trustworthy.
If the underlying setup is benign, computation results can be considered likewise.
However, these attestation mechanisms rely on hardware trust-anchors and a chain
of trust which may not be available or may be bypassed.

A different approach is addressed by verifiable computation methods. These aim
to solve the integrity problem by producing cryptographic proofs of correct
execution along with the actual computation result. As the proof generation is
embedded into the algorithm, a proof is always tied to a certain result and can
only be generated if the computation was executed correctly. Using verifiable
computation, it is possible to remove the underlying hard- and software from a
serviced trusted computing base and thus widen the range of target devices and
strengthen the confidence in computation results.

This seminar focuses on concepts and applications of state-of-the-art verifiable
computation methods such as:

- Probabilistically Checkable Proofs
- Interactive Proofs
- Argument Systems
- Bulletproofs
- Verifiable State Machines


- IN0015 Diskrete Strukturen
- MA0901 Lineare Algebra für Informatik
- IN0018 Diskrete Wahrscheinlichkeitstheorie
- preferable: IN0042 IT Sicherheit


Capacity: The seminar is suitable for up to 9  students.