Google file system Q: (be a spy!) We've read a paper by one of the three authors in the GFS paper in a previous class? Which one? What's the major difference between that paper and this? Why the change in design? Q: Does GFS offer the same close-to-open consistency as NFS? If so, how? If not, why not? Chubby Q: What was Chubby designed for? Give some usage scenarios. Q: We have built a lock service in the labs too. Comment on the difference in designs of our lock service and Chubby in terms of: * purpose of the lock service * lock granularity * failure tolerance (of clients and lock server itself) * Would Chubby be helpful for use in the lab?