Multiple choice Quiz Flashcards
q3. Which of these are abstractions that will be covered in the course?
Shared memory
Consensus
Reliable broadcast
q3. Communication between components in our programming model can be done by:
Sending and receiving messages (only)
The event-based programming model can only communicate using events/messages, and doesn’t have shared variables.
q3. Events in our model can be:
Messages
Timers
Conditions (e.g. x == 5 && y < 9)
q3. The interface for a channel component have the following events:
A deliver indication event
A send request event
The events need to have the correct type (request or indication).
q3. Which of the following should be part of a service specifications?
Service interface (API) The required system model Correctness properties
The specification only describes “what”, not “how”.
q3. The service API describes
Responses (indication events)
Requests (request events)
q3. A safety property states that
Nothing bad ever happens
q3. A liveness property states that
Something good eventually happens
q3. Is it the case that any property can be described as a conjunction of a safety property and a liveness property?
True
This has been proven formally
q3. Safety can only be satisfied in infinite time
Safety can be violated in finite time
True
q3. Liveness can only be violated in infinite time
Liveness can be satisfied in finite time
True
q3. If it doesn’t fail in the (entire) execution
A correct process is one which never fails at any point in the execution.
True
q3. In the crash-stop model, processes are not allowed to recover
True
q3. In the crash-recovery model, a process is faulty if
It crashes and never recovers
It crashes and recovers infinitely often
q3. Which of the following models are special cases of the crash-recovery model?
Omission
Crash-stop
q3. The channel model…
Describes characteristics/guarantees of the network
q3. When using a fair-loss link, if a message m is sent infinitely often by pi to pj, and neither crash, what is guaranteed to happen?
m is delivered an infinite number of times by pj
q3. When using a fair-loss link, if a message m is sent 10 times by pi to pj, how many times may the message be delivered?
A message cannot be duplicated infinitely many times.
q3. When using a fair-loss link, if a message m has not been sent by any process to pj, is pj allowed to deliver m?
No message is delivered unless it was sent.
q3. When using stubborn links,if a node pi sends a message m to a correct node pj, and pi does not crash, how many times does pj deliver m?
An infinite number of times
q3. Is it possible to implement the stubborn linkserviceusing the fair-loss link service?
True
q3. When using perfect links, if a message m is sent by pi to pj, andneither pi nor pj crashes, is pj guaranteed to deliver m?
Yes, eventually (unbounded time)
q3. When using perfect links, if a message m is sent by pi to pj, how many times is the message allowed to be delivered?
At most once
q3. Which kind of system is described by the following statement: my server always processes requests within one week.
Synchronous system
There is a known bound on node processing delay.
q3. Which kind of system is described by the following statement: my server processes requests within one week when it is running, and it will eventually be running for at least a week, I just don’t know when that will be.
Partially synchronous system
q3. Safety in distributed systems describes:
a property that states that the system is robust to malicious behavior
a property that states that an event that is applicable is immediately applied
a property that holds for every execution in some configuration
a property that holds in every reachable configuration
q3. Liveness in distributed systems describes:
a property that ensures low time complexity
a property that states that an event that is applicable infinitely many times is eventually applied
a property that holds for every execution in some configuration
a property that states that an event that is applicable infinitely many times is applied infinitely many times
Select all cases when process p is correct, assuming the crash-recovery model. CORRECT
p crashes and recovers once every minute for the first year but after that never crashes again
p never crashes
A new and highly advanced computer system has been installed and is running. The system guarantees that processes can send messages to each other in less than 10 microseconds. But the system needs to run an automatic tuning procedure before this 10 microsecond-guarantee starts to hold, and during this procedure the communication delays are unbounded. No one knows how long time the tuning procedure will take.
Which model best describes this system?
Partially synchronous