How can an attacker abuse a distributed system?

Question

How can an attacker take advantage of a world where state is eventually consistent and vector clocks are used to settle disagreements?

For example, what are the security concerns of using a database like Cassandra to settle finical transactions?

Is anyone aware of exploit code that is leveraging a flaw in a distributed system?

Answer 1

This post is going to serve as a repository of references possibly related to the problem. Once there is a substantial answer, please copy the contents into it and ping me or the mods. Please feel free to edit this answer to refine the list.

As I've said, CiteSeerX yields nothing useful (its search system isn't exactly top-notch). However, this query on arXiv gives the following results:

• Bing-Rong Lin, Ye Wang, Shantanu Rane. On the Benefits of Sampling in Privacy Preserving Statistical Analysis on Distributed Databases.
• Wenbing Zhao. A Byzantine Fault Tolerant Distributed Commit Protocol
• Petros Maniatis, Mary Baker. Secure History Preservation Through Timeline Entanglement

Search through Google Scholar:

• P.Liu, J.Jing. Architectures for Self-Healing Databases under Cyber Attacks
• B.Bhargava, L.Lilien. Vulnerabilities and threats in distributed systems (apparently a neat overview)