Cloud computing is an attempt to stretch the operation boundaries dynamically on-demand which is not the case with traditional distributed system. This unprecedented nature of cloud though offers various benefits it also has a downside when it comes to fault tolerance since most of the faults experienced by the cloud are unpredictable and thus Byzantine in nature. Byzantine faults in cloud comp...

Cooperative, peer-to-peer (P2P) services—distributed systems consisting of participants from multiple administrative domains (MAD)—must deal with the threat of arbitrary (Byzantine) failures while incentivizing the cooperation of potentially selfish (rational) nodes that such services rely on to function. This paper investigates how to specify conditions (i.e., a solution concept) for rational ...

Clock synchronization is a very fundamental task in distributed system. It thus makes sense to require an underlying clock synchronization mechanism to be highly fault-tolerant. A self-stabilizing algorithm seeks to attain synchronization once lost; a Byzantine algorithm assumes synchronization is never lost and focuses on containing the influence of the permanent presence of faulty nodes. Ther...

This thesis presents a scheme that achieves self-stabilizing Byzantine digital clock synchronization assuming a “synchronous” system. This synchronous system is established by the assumption of a common external “beat” delivered with a regularity in the order of the network message delay, thus enabling the nodes to execute in lock-step. The system can be subjected to severe transient failures w...

We present a scheme that achieves self-stabilizing Byzantine digital clock synchronization assuming a “synchronous” system. This synchronous system is established by the assumption of a common external “beat” delivered with a regularity in the order of the network message delay, thus enabling the nodes to execute in lock-step. The system can be subjected to severe transient failures with a perm...

The growth of data, the need for scalability and the complexity of models used in modern machine learning calls for distributed implementations. Yet, as of today, distributed machine learning frameworks have largely ignored the possibility of arbitrary (i.e., Byzantine) failures. In this paper, we study the robustness to Byzantine failures at the fundamental level of stochastic gradient descent...