Вы находитесь на странице: 1из 3

Reg. No.

Question Paper Code :

31106

Common to M.E. Computer Science and Engineering/M.E. Mobile and Pervasive Computing Second Semester

241205 ADVANCED OPERATING SYSTEMS (Regulation 2010) Time : Three hours

50

Answer ALL questions.

1. 2.

Mention the functions of operating systems. Why does the interrupt disable method to achieve mutual exclusion not work for multiprocessor systems? State the purpose of control site in completely centralized deadlock detection. How does Lamports algorithm guarantee mutual exclusion? Sender-initiated algorithms cause system instability at high system loads. Predict analytically at what system load the instability will occur. Assume probe limit is 5, average service requirement of a task is 1 sec, overhead incurred by a processor to poll or to reply to a poll is 3 msec. Why is mapping function required in the central server algorithm? Differentiate between forward-error and backward-error recovery. Define global atomicity. Give an example.

3. 4. 5.

75

07 7

6. 7. 8. 9.

List out the issues in preprocessor scheduling that causes performance degradation in multiprocessor systems.

10.

Write the difference between tightly coupled system and loosely coupled system with an example.

50

77

PART A (10 2 = 20 marks)

77

M.E./M.Tech. DEGREE EXAMINATION, JUNE 2011.

50
Maximum : 100 marks

PART B (5 16 = 80 marks) 11. (a) (i) (ii) Describe Bankers algorithm supporting example. for deadlock

50
avoidance

Or (b)

(i) (ii) 12. (a)

Producer Consumer Problem Readers Writers Problem.

50
Or 2

Give a solution for the following synchronization problems using semaphores. (8) (8)

Discuss in detail about the communication models that provide communication primitives. (16)

(b)

75

07 7

(i)

(ii)

(iii) Determine the number of messages exchanged to detect a deadlock in the worst case. (4) (iv) Determine the saving (in percentage) in the average number of messages exchanged under this message traffic reduction scheme, as compared to when no such is used. (4)

50

Consider the following scheme to reduce message traffic in distributed deadlock detection: Transactions are assigned unique priorities, and an antagonistic conflict occurs when a transaction waits for a data object that is locked by a lower priority transaction. Deadlock detection is initiated only when an antagonistic conflict occurs. When a waiting transaction receives a probe that is initiated by a lower priority transaction, the probe is discarded. Determine the number of messages exchanged to detect a deadlock in the best case. (4)

Determine the number of messages exchanged to detect a deadlock in the average case. (4)

77

77

Consider a computer system which has four identical units of a resource R. There are three processes each with a maximum claim of two units of resource R. Processes can request these resources in anyway, that is, two in one shot or one by one. The system always satisfies a request for a resource if enough resources are available. If the process doesnt request any other kind of resource, show that the system never deadlocks. (8)

7
with (8)

31106

13.

(a)

(i) (ii)

Predict the performance of the receiver-initiated load sharing algorithm when the entire system workload is generated at only a few nodes in the system instead of equally at all nodes in the system. (8)

(b)

Describe about the Distributed Shared Memory. Explain the algorithms that are developed for implementing the DSM. (16) (i) (ii) Define livelocks. What is the difference between a deadlock and livelock? (6) Explain the Asynchronous Checkpointing Recovery method and give an example where the recovery algorithm will need to execute for [N] iterations where [N] is the number of processors in the system. (10)

14.

(a)

(b)

Design a decentralized two-phase commit protocol,

(ii)

Which uses only O( N ) messages where N is the number of sites in the system. (8) Can the performance of a multiprocessor system with two processors be worse than the performance of a uniprocessor (with an identical CPU)? Explain the answer with an example. (8)

15.

(a)

(i)

(ii)

50

A task consists of several subtasks. If these communicate synchronously with each other frequently, which scheduling policy will suit for this scenario and why? (8) Or

07 7

(b)

Discuss in detail about the lightweight process and its different levels. Write its advantages, disadvantages and performance of supporting lightweight process at these levels. (16)

75

77
3

(i)

Where no site is designated to be a coordinator

50
Or

77

Or

50

With a neat diagram explain the architecture and data access of a distributed file system. (8)

(8)

31106

Вам также может понравиться