http://www.makaut.
com
Name : ……………………………………………………………
Roll No. : ………………………………………………………..
Invigilator’s Signature : ………………………………………..
CS/M.Tech (CSE)/SEM-2/MCS-203/2010
2010
DISTRIBUTED COMPUTER SYSTEMS
Time Allotted : 3 Hours Full Marks : 70
The figures in the margin indicate full marks.
Candidates are required to give their answers in their own words
as far as practicable.
Answer any five questions. 5 × 14 = 70
1. a) What are the different types of hardware architectures
on which a distributed system can be built ? Explain
their main features & the types of O.S. that utilize these
hardwares. 4
b) What is the Cache coherence problem ? Explain a
directory based protocol which provides a solution to
the problem. 2+4
c) In case of NORMA architecture, what strategies are
taken to improve performance when access is made to
i) read only page
ii) writable page ? 4
30068 (M.TECH) [ Turn over
CS/M.Tech (CSE)/SEM-2/MCS-203/2010
2. a) What are the assumptions involved for ensuring a
sequentially consistent execution of a program ?
Compare with strict consistency & explain with the help
of a program code example. 6
b) How is it possible to enforce strict consistency on
program codes which are basically sequentially
consistent ? 4
c) In most implementations of Release Consistency in DSM
system shared variables are synchronized on Release
but not on Acquire. Why is Acquire needs at all ? 4
3. a) Why is Logical clock in a distributed system more
relevant from the point of view of Event ordering ? Give
two examples of situation where synchronization on
Physical clock time is essential. 2+2
b) State the principles behind Lamport’s Algorithm for
clock synchronization. 4
c) What are the drawbacks of a Centralized Clock
synchronization algorithm ? Give an example of
Decentralized Clock synchronization algorithm. How
does it improve upon the problems inherent in the
Centralized algorithm ? 6
30068 (M.TECH) 2
CS/M.Tech (CSE)/SEM-2/MCS-203/2010
4. a) In a distributed system running the same kernel image,
what is the generic methodology for controlling access to
the critical region ? Describe any algorithm. 2+4
b) What are the usual load measurement strategies in a
distributed system ? Describe a processor allocation
algorithm where co-ordinator failure does not form a
bottleneck. 2+6
5. a) What are the necessities of properties – Atomicity,
Consistency, Isolation & Durability in distributed
transaction processing ? Explain with the help of
examples. 8
b) A process with transaction timestamp 50 needs a
resource held by a process with transaction timestamp
100. What happens in
i) Wait – Die
ii) Wound – Wait. 3+3
6. a) Describe the situations in which an Election algorithm
is used. Also describe any one of Election algorithms. 7
b) What is the type of architecture which can support a
DSM ? Describe the three main models followed for the
implementation of DSM. 1+6
30068 (M.TECH) 3 [ Turn over
CS/M.Tech (CSE)/SEM-2/MCS-203/2010
7. a) If in a remote procedure call – read (fd, buf, nbytes)
where buf represents a pointer to an array is called by a
Client program, how is the parameter passing done as
the memory architecture for the client & the server may
be different ? 6
b) Explain the relative advantages and disadvantages of
stateful and stateless servers. 4
c) ‘‘VNODE’’ is a generalization of ‘‘INODE’’ in the Unix File
System architecture. Explain the statement in the
context of Network File System. 4
8. Write short notes on any two of the following : 2×7
a) Detection of Byzantine System Faults.
b) Transaction Rollback using Intentions List & two phase
commit protocol where one subordinate process is not
prepared to commit.
c) Application of Voting algorithm in a Replicated File
System.
30068 (M.TECH) 4
