Failure detectors are fundamental building blocks in distributed systems. Multi-node failure detectors, where the detector is tasked with monitoring N other nodes, play a critical...
The large size and high complexity of securitysensitive applications and systems software is a primary cause for their poor testability and high vulnerability. One approach to all...
Distributed systems are hard to build, profile, debug, and test. Monitoring a distributed system – to detect and analyze bugs, test for regressions, identify fault-tolerance pr...
The database tier of dynamic content servers at large Internet sites is typically hosted on centralized and expensive hardware. Recently, research prototypes have proposed using d...
Abstract. We analyze the expected cost of a greedy active learning algorithm. Our analysis extends previous work to a more general setting in which different queries have differe...