Give two reasons synchronizing clocks is difficult in a distributed system.
Briefly describe how Critian's algorithm works.
Describe what happens in the "bully" election algorithm when a server/leader goes down. In a practical system, what aspect of the bully algorithm may be difficult?
What is the atomic nature in regard to transactions? Why is this more difficult in a distributed system than in a centralized system?
How is XML different than HTML?
What is the difference between a persistent connection and a parallel request? Describe system characteristics where you would want to use each.
Then, why would you want to use strong cache consistency? Why would you want to use weak cache consistency?
Describe one weak cache consistency heuristic. When might it fail?
Where does the SSL layer reside in the network stack?