Total No.
of printed pages = 4
CSE 1817 OE 31
Roll No. of candidate j I I I I. I I I I I /
2021
[Link]. 7th Semester End-Term Examination
CSE
DISTRIBUTED SYSTEMS
(New Regulation w.e.f. 2017 - 18 & New Syllabus w.e.f. 2018 - 19)
[Link] Papers
Full Marks - 70 Time - Three hours
The figures in the margin indicate full marks
for the questions:
Answer question No. 1 and any four from the rest.
1. Answer the following (MCQ/ Fill in the blanks).: (10 X 1 = 10)
(i) In a distributed system, each processor has its own
(a) Clock
(b) Local memory
� Both local memory and clock
(d) None of the above·
(ii) What is the common problem found in Distributed systems?
(a) Process Synchronization
(b). Communication synchronization
_{_c) · Deadlock problem
(d) Power faiiure
__,
(iii) The characteristics of the tightly coupled system. are
- (a) differe�t clock
. ,
(b) use communication links
!};}' same clock
(d) distributed systems
[Turn over
(iv) In messoge-possing systems, o mossage-pHAsing facility provides at lcAAt
two operations
(a) send (message) and delete (message)
(b) delete (message) and receive (message)
(c) write (message) and rend (message)
send (message) and receive (message)
�
(v) Single-unit request model is a particular case of the __ request model.
(vi) Lamport's non-token-based algorithm can be optimized by suppressing
___ -messages in certain situations.
(vii) ___ command creates and starts a remote object registry on the specified
port on the current host.
[Link] Papers
( viii) When any single application function or component fails, then the entire
application goes down
(a) Microservices
(b) Macroservices
J.c) Monolithic
(d) All of the above
(ix) In case of failure, a new transaction coordinator can be elected by
(a) Bully algorithm
'-/'
(b) Ring algorithm
(c) Both bully and ring algorithm
(d) None of the mentioned
(x) The deadlock handling in distributed system 1s highly focused towards
deadlock detection methods. Because
( a) Once a cycle is formed in the state graph, it persists until it is detected
and broken.
.. _(b,) Deadlock avoidance is impractical in distributed systems.
(c) Deadlock prevention 1s inefficient because it decreases system
concurrency
._# All of the above
CSE 1817 OE31 2
2. JP')· "Th o 'l\\18111 ordering of nwARug A hPlpR to dr vnlnp the cJu�tribuf/Jd
nll:\ rithm," Di1-1 \IBB thiR stnt m nl by •xploining with o rel •vanf profoc(JJ.(7)
(b) Dis ·u s the inhor •nt, limilationa of diAtributcc.l syAt mA and their impr.1ct on
th design und development of the syatom. 8)
3. (a) Illu trule the following terminologies:
(i) inconsistency between two local stateA
(ii) strongly consistent global state (5)
Qp. The following figure shows fot�r processes (Pl, P2, P3, P4) with eveni8 a,b,c
[Link] Papers
u and message events communicating between them. Considering all initial
logical clock values as 0.
List the Vector Clock timestamps for each event shown in the figure
.jv below. Provide timestamps for each labeled event.
(ii) List the events that are concurrent a11d explain why? (7 +3 =10)
-:!;.�
Pl
P2
P3
P4
\ 4. (a) Explain the Chandy-Mishra-Haas's AND request model in a distributed
) deadlock detection. · (8)
\
� What are two conditions that mu�t be satisfied by a correct deadlock
detection algorithm? Discuss in detail (7)
5. (a) Differentiate between flat and nested distributed transactions with the help
of diagram. (5)
(b) What are the different timestamp-based locking algorithms for concurrency
control used to resolve
-
the conflicts,in a distributed database systems. (10)
3 [Turn over
CSE 1817 OE 31
1 � .. how thnt tho rriti nl. cl ion iA ncco�Acd nccorclin� to th,, inrr<'Mun,� ord,•r ,,r
timl' �tnmp, in th Jlccnrt-ARrawnln nl�orithm. (7)
(b) Mnekawa's algorithm is prone to deAdlock. Explain the dctAil of deadlock
handling steps using three lypes of messages. Whal is the maxim um
number of messages required per critical section execution in this case?
(7+ 1=8)
7. (l\} _.... Discuss and compare the performance of distributed token and non-token
T based mutual exclusion algorithms specially under light load and heavy Jo::irl
conditions. (8)
'What is distributed shared memory? Illustrate some ::idvantfiges of it.
(3+4= 7)
[Link] Papers
4
CSE 1817 OE 31