Content sharing is a popular usage of peer-to-peer systems for its inherent scalability and low cost of maintenance. In this paper, we leverage this nature of peer-to-peer systems to tackle a new problem: automatic misconfigu-ration troubleshooting. In this setting, machine configurations from the peers are "shared" to diagnose the misconfigurations on a sick machine. A key challenge for such a troubleshooting system is privacy preservation. To this end, we construct Friends Troubleshooting Network (FTN), a peer-to-peer overlay network, where the links between peer machines reflect the friendship of their owners. To preserve privacy, we use historyless and futureless random-walk in the FTN, during which search along with parameter aggregation are carried out for the purpose of troubleshooting. Many of our techniques can be applied to other application scenarios that require privacy-preserving distributed computing and information aggregation. We have also identified a number of open challenges that remain to be addressed.
展开▼
