[Back]

B. Körbel:
"A Blockchain-based Computation Offloading Approach with Result Verification";
Supervisor: S. Schulte, M. Sigwart, P. Frauenthaler; Institute of Information Systems Engineering, Distributed Systems Group, 2020; final examination: 2021-01-29.

Computation offloading refers to the outsourcing of tasks to providers with vast computational resources in exchange for a fee. It can be an enabling technique for devices with limited computational resources, e.g., mobile or IoT-devices. These resource-constrained devices can benefit from computation offloading, e.g., by a reduced application runtime and battery savings.
Conventionally, computation offloading takes place in the cloud, i.e., in data centers with nearly endless computational resources. Usually, all layers of the cloud such as the assignment of resources and payment processing are controlled by a single provider. Because of high cost for provisioning and maintaining cloud infrastructure, the cloud provider market is dominated by only a few big organizations such as Amazon, Microsoft, and Google. For small to medium-sized companies it is difficult to enter the market.
Blockchain technology provides an alternative to traditional cloud providers. Organized as peer-to-peer network, the blockchain manages data and programs in a decentralized manner since all participants execute and verify all state changes. Furthermore, blockchain
technology allows participants to freely join and leave the network. As such, blockchain technology is a promising solution to realize a decentralized and open computation offloading platform. However, carrying out resource-intensive computations on blockchains
is rather expensive or in some cases just not possible. To tackle this problem, computations can be executed off the blockchain. However, if the task is computed off-chain, the correctness of the computation is not verified by the peer-to-peer network. In other words, the resource provider has to be trusted.
In this thesis, we develop a blockchain-based computation offloading approach that allows arbitrary providers to rent out their computational resources for a fee. We compare multiple solutions for executing tasks off-chain without having to trust the resource
provider carrying out the computation. As such, the proposed solution provides an open alternative to traditional cloud providers.

Computation Offloading / Blockchain / Zero Knowledge Proofs