Exercises for lesson 2, TDDD82 System softare module, Linköping University
Q1:
The figure below shows the resource allocation graph for five processes and thre resource
types (R1,R2, and R3). Assume that this represents the system state at a specific moment in
time.
(a) Is the system currently in a deadlock state?
(b) Assume now that the arrows in the graph also represent the maximum number of
resources that a process can request. Is there a sequence of resource assignments
that allow all processes to finish their executions without the system ending up in
a deadlock?
P2
P1
P3
R1
R3
P5
R2
P4
Q2:
Describe how synchronization between processes can be accomplished with the help of the
monitor concept. What mechanisms are there that support the application programmer?
Exemplify this method by describing a system where two processes produce sensor values
and a third process combines those values and outputs a result. Assume that the sensors
produce data at different rates and that the combination process should use a pair of elements
(one from each sensor) every time.
Q3:
Argue how parallelism occurs in computer systems and the use of modelling it.
Exercises for lesson 2, TDDD82 System softare module, Linköping University
Q4:
(a) The folloing pseudocode uses the semaphores wait and signal with the purpose of
achieving mutual exclusion of a shared resource. Does this implementation achieve
the objective? If so, explain why. If not, provide a trace from the execution of the
processes leading to an unwanted result.
process P1
signal(mutex)
critical_section_1
signal(mutex)
non_critical_section_1
end
Process P2
wait(mutex)
critical_section_2
signal(mutex)
non_critical_section_2
end
(b) Explain what happens in the system if
a.
P1 crashes while in its non-critical section,
(non_critical_section_1).
b. P2 crashes while in its non-critical section.
Q5:
Consider two communication services in asynchronous distributed systems. Messages that are
treated by service A can be lost, duplicated, or delayed and checksum is only used on the
header field. For service B, messages can be lost, delayed or delivered in a too high rate to be
processed by the receiver, but always with the correct content. Describe the fault model for
each service and explain how the fault may affect validity and integrity properties. Is service
B a reliable communication service?
Q6:
Consider a simple Internet server that respons to client requests without querying other
servers. Explain why it is not poessible to set a bound on the time the services takes to
respond to the client. What would be required to guarantee that the server responds within a
specified time interval? Is your proposal reasonable?
Q7:
Explain the difference between permanent, transient and intermittent faults, and give one
example for each kind.
Q8:
Consider a replicated service that is implemented using three processes. Assume that every
node has a MTTF = 5 days, and that it takes 4 hours to repair a failure. What is the
availability of the replicated service?
Q9:
Explain with the help of an example the concept of consistency in connection with state
transfer in distributed systems.