Birla Institute of Technology & Science, Pilani Work-Integrated Learning Programmes Division Second Semester 2009-2010 Comprehensive Examination
(EC-2 Regular) Course No. Course Title Nature of Exam Weightage Duration Date of Exam Note:
1. 2. 3. 4.
: CS ZG623 : ADVANCED OPERATING SYSTEMS : Open Book : 60% : 3 Hours : 04/04/2010 (AN)
No. of Pages =2 No. of Questions = 10
Please follow all the Instructions to Candidates given on the cover page of the answer book. All parts of a question should be answered consecutively. Each answer should start from a fresh page. Mobile phones and computers of any kind should not be used inside the examination hall. Use of any unfair means will result in severe disciplinary action.
Q.1 (a).
What is the difference between logical time and real time? Why is not real time being used for all distributed systems? [3] What kind of problems are Local Clocks useful for? What is needed in order to use the local clock as a global clock? [3] Receiver initiated load sharing algorithms are better than Sender initiated ones in high load situations. Discuss the performance of the receiver-initiated load sharing algorithms when the entire system workload is generated at only a few nodes in the system instead of equally at all the nodes in the system. [3] Assume that several users are using a single copy of an editor program to edit their files. This editor program is accessing the data areas of different users by using a capability. If Access Matrix Model is to be used, explain how you can implement this using capability based addressing? [3] The Sun NFS client module caches the results of read, write, getattr, lookup, and readdir operations in order to reduce the number of requests transmitted to servers hence making it a scalable design. Describe how the client ensures that the data are up-to-date. [3] In a distributed system, it is argued that UNIX semantics for file sharing is not practical, because of performance reasons, as it needs all file requests to go to a single server and no client caching is allowed. Suggest and explain another semantics model that would accommodate multiple file servers, client caching, and most of the common operations allowed on a file. Also explain how your model can handle the situation where two or more clients are simultaneously caching and modifying the same file. [3]
Q.1 (b).
Q.2 (a).
Q.2 (b).
Q.3 (a).
Q.3 (b).
Q.4. In the following figure, a sequentially consistent memory allows how many possible statement interleaving? List them all. [6]
Process P1 x=1; if (y==0) kill (P2); Process P2 y=1; if (x==0) kill (P1);
CS ZG623 (EC-2 Regular) Second Semester 2009-2010 3
Page 1 of
CS ZG623 (EC-2 Regular) Second Semester 2009-2010
Page 2
Q.5. Assume that BITS Computer Science & Information Systems department has a multi-user computer, few workstations and a number of terminals. The Unit Chief wants to extend the computing resources. In place of installing another multi-user computer terminal, he decides to go for having a distributed system. What are the design issues involved in this new proposal to achieve the goal? Discuss. [6] Q.6. Is node P allowed to send messages related to the application (as opposed to message that are part of the checkpoint algorithm itself) immediately after having taken a tentative checkpoint? If so, explain why this is permitted. If not, give an example that illustrates why P should not be allowed to do so and state a condition that defines when P may resume sending messages related to the application. [6] Q.7. In the following Wait for Graph, use Obermarcks Path pushing algorithm to build local WFGs at each site, and also show the Path (Cycles) that will be transmitted by each site to other dependent sites to detect a deadlock if any. [6]
P1 P2 P4 P5
S2
S1
P3
P6
P7 P9
P8
S3
Q.8 (a).
The receiver initiated distributed scheduling algorithms assumed task pre-emption at the sender site. Devise an approach, which is receiver-initiated but non pre-emptive. Discuss the tradeoffs of your approach with the receiver initiated pre-emptive one. [4] Java and other languages support exceptions, which are raised when an error occurs. How would you implement exceptions in RPCs? Explain. [2] Discuss how well the load sharing implementations (System V, Condor, and Stealth) satisfy the scalability criterion. [3] Figure below shows a distributed system with three processes P1, P2, and P3. Each process has different send and/ or receive events. Find out all the happened before relations. [3] Space
Q.8 (b).
Q.9 (a).
Q.9 (b).
Q.10 (a).
P1 P2 P3 Time What is total ordering of events? Why is it required? How is it achieved? Explain with an example. [3] Naming is an important issue to be considered while addressing the design of distributed file systems. What are the different approaches to name files in Distributed Systems? Which one you would like to prefer as a designer and why? Explain. [3] Page 2 of
Q.10 (b).
CS ZG623 (EC-2 Regular) Second Semester 2009-2010 3
*********
CS ZG623 (EC-2 Regular) Second Semester 2009-2010 3
Page 3 of
