An agreement on a faulty system cannot be reached in which scenario?
Unbounded, ordered, asynchronous, multicast
Unbounded, ordered, syncronous, multicast
Unbounded, ordered, asynchronous, unicast
Unordered, bounded, asynchronous, multicast
retransmission if packet lost is an example of
backward recovery
forward recovery
retry recovery
automated recovery
Which of the following is a drawback of using sequence numbers to prevent duplicates when retrying?
Results have to be stored for some time
Only works for idempotent requests
Sequence number can never reset and therefore eventually gets too big to store
Blocks intermediate processes from making calls while the failed call is reissued
What does the following describe "construct correct state from error state"
extrapolation
retransmission
What type of logging does "roll each process back to state where does not depend on delivery of message some process(es) do not yet have in stable storage" describe?
Optimistic logging
Pessimistic logging
Reliable logging
Transaction logging
Which of the following describes "virtual synchronous multicast"
sender crashes, deliver to all rest of processes or none
sender sends requests to three receivers and uses triple modular redundancy to detect errors
If sender has crashed by the time the receiver responds, the receiver selects a new node to deliver the response to
Multiple recievers are guaranteed to receive the message at exactly the same time
Which of the following does this describe "requires that an operation is performed by each member of a process group or none at all"?
distributed commit
atomic transactions
coordinated processing
synchronous computation
"server fails to receive incoming messages" is what type of failure?
Crash failure
Omission failure
Response failure
Timing failure
In which of the following M -> P scenarios will a crash result in a duplicate if calls are always reissued?
MC(P)
C(MP)
MPC
Which of the following does not offer Total-ordered Delivery?
Causal atomic multicast
FIFO atomic multicast
Atomic multicast
Reliable multicast
In which of the following P -> M scenarios will be okay if calls are reissued only when ACKed?
PMC
PC(M)
C(PM)
Which of the following is false about RPC?
Communication failures such as lost messages means RPC can't be completely transparent
It can be difficult to determine the difference between a failed message and a delayed reply
Reissuing RPC calls can have undesirable consequences
An RPC call can always be reissued because changes resulting from RPC call must be acknowledged
Is a hit counter on a website idempotent?
What is the minimum number of nodes required to detect 2 faulty nodes in Byzantine failures?
5
3
7
4
