COMP28112
Two hours
Question ONE is COMPULSORY
UNIVERSITY OF MANCHESTER
SCHOOL OF COMPUTER SCIENCE
Distributed Computing
Date: Friday 20th May 2016
Time: 09:45 - 11:45
Please answer Question ONE and also TWO other Questions
from the remaining THREE Questions provided
This is a CLOSED book examination
The use of electronic calculators is NOT permitted
[PTO]
, COMP28112
1. Compulsory
a) What is a Java servlet? (2 marks)
b) What is the main assumption on which Cristian’s clock synchronisation
algorithm is based? (2 marks)
c) Explain briefly the difference between a name server and a directory server.
(2 marks)
d) Explain briefly what the relevance of the Byzantine generals problem is for
distributed computing systems. (2 marks)
e) Why is it practically impossible to achieve strict consistency in a distributed
system? (2 marks)
f) Explain what copy-restore is and why it is used in RPC. (2 marks)
g) What is the worst-case scenario for the number of messages in the Bully
algorithm? How many messages do you expect in this case? (2 marks)
h) What is the key difference between caching and replication? (2 marks)
i) Explain briefly what the use of a checkpoint is in the context of distributed
transactions. (2 marks)
j) In the context of lab exercise 2, what would you do to launch a denial of
service attack against the server? (2 marks)
Page 2 of 6
Two hours
Question ONE is COMPULSORY
UNIVERSITY OF MANCHESTER
SCHOOL OF COMPUTER SCIENCE
Distributed Computing
Date: Friday 20th May 2016
Time: 09:45 - 11:45
Please answer Question ONE and also TWO other Questions
from the remaining THREE Questions provided
This is a CLOSED book examination
The use of electronic calculators is NOT permitted
[PTO]
, COMP28112
1. Compulsory
a) What is a Java servlet? (2 marks)
b) What is the main assumption on which Cristian’s clock synchronisation
algorithm is based? (2 marks)
c) Explain briefly the difference between a name server and a directory server.
(2 marks)
d) Explain briefly what the relevance of the Byzantine generals problem is for
distributed computing systems. (2 marks)
e) Why is it practically impossible to achieve strict consistency in a distributed
system? (2 marks)
f) Explain what copy-restore is and why it is used in RPC. (2 marks)
g) What is the worst-case scenario for the number of messages in the Bully
algorithm? How many messages do you expect in this case? (2 marks)
h) What is the key difference between caching and replication? (2 marks)
i) Explain briefly what the use of a checkpoint is in the context of distributed
transactions. (2 marks)
j) In the context of lab exercise 2, what would you do to launch a denial of
service attack against the server? (2 marks)
Page 2 of 6
