Category: Uncategorized

Global Memory Systems The context for this question is the same as the previous question. [5 points] Consider a 4 Node GMS system (A, B, C, D). Node A needs to evict pages P1 (age=15) and P2 (age=28). In the current epoch, the parameters are as follows: {“minAge”:25, “”:{“A”:0.2, ”B”:0.3, ”C”:0.1, ”D”:0.4}} c) [2 points] Mention any additional work that the node A may have to do when the pages are evicted.  What is the justification for any such additional work?

Potpourri The context for this question is the same as the previous question. [6 points]  Your team is experimenting with different RPC communication methods for designing a distributed system. Two design choices are being considered- Design A involves LRPC-style mechanism where the client and server communicate through shared memory in an SMP. Design B involves traditional RPC where the client and server communicate across the LAN as discussed in the Thekkath & Levy paper. c) [1 point] How can you reduce the number of context switches in Design A even further?

Global Memory Systems The context for this question is the same as the previous question. [5 points] Consider a GMS system with 10 Nodes ‘A’ to ‘J’. A process ‘x’ on Node ‘A’ intends to read-share a page that is  currently actively used only by a process ‘y’ on Node ‘B’.When Node A’s GMS mechanism initiates the access, indicate on which Node(s) each of the following data structures  are accessed.Assume no co-location constraints for any of the data structures b) [1 point] GCD

Spring OS [3 points] Your company has two teams developing Spring services that must interoperate. One team is implementing their service in C, while the other team is using C++. What artifacts must each team produce, so that these services, written in different languages, can communicate seamlessly within the Spring system? Briefly explain how these artifacts will be generated?

Distributed Shared Memory The context for this question is the same as the previous question. [8 points] In Treadmarks DSM system the following critical section is executed at a node N1: Lock(L1); // Assume X is the only page that been modified up to this point by prior acquirers of L1     Write to Page X;      // Assume that the page is not present at this node;      // Assume that there are three diff files for page X      // named Xd2, Xd3, and Xd4 in nodes N2, N3, and N4      // respectively.      // Assume the sync causality for the lock L1 is       // N3 -> N2 -> N1 (i.e, this is the order of lock     // acquisition).Unlock (L1); b) [2 points] What additional actions would be carried out by Treadmarks at Node 1 before the critical section  is executed by N1?

Spring OS The context for this question is the same as the previous question. [8 points] Your coworker is debugging a distributed payment processing service built on Spring OS. Currently, the client process hangs when ‘invoice.submit()’ is called. The client process runs on Node A, but the invoice service object lives on Node B on a LAN. Your coworker would like to better understand the runtime plumbing so they can trace, reproduce, and fix the bug. c) [3 points] At run time, explain the flow of the invocation from the client to the invoice service.

Distributed File Systems Background: The Google File System was designed a few years after xFS and is heavily influenced by xFS. GFS also separates its metadata management logic from its data storage. It uses a single “Master” server to store all file system metadata. It uses many “Chunkservers” to store actual data blocks. A client first contacts the single Master to get metadata and then contacts the Chunkservers directly to get data. Give one pro and one con of the GFS approach to metadata management.

Distributed Shared Memory The context for this question is the same as the previous question. [8 points] In Treadmarks DSM system the following critical section is executed at a node N1: Lock(L1); // Assume X is the only page that been modified up to this point by prior acquirers of L1     Write to Page X;      // Assume that the page is not present at this node;      // Assume that there are three diff files for page X      // named Xd2, Xd3, and Xd4 in nodes N2, N3, and N4      // respectively.      // Assume the sync causality for the lock L1 is       // N3 -> N2 -> N1 (i.e, this is the order of lock     // acquisition).Unlock (L1); c) [2 points] Upon exiting the critical section what action would be carried out by Treadmarks at Node N1?

Java RMI [3 points] The Java RMI system consists of three main layers: the stub/skeleton layer, the remote reference layer, and the transport layer. For each layer, give one succinct bullet explaining each of these layers.

Lamport’s ME Algorithm The context for this question is the same as the previous question. [6 points] Consider a distributed system with 3 nodes (P1, P2, P3) implementing Lamport’s mutual exclusion algorithm, as shown in the Figure. The boxes represent requests made by a node at a particular timestamp. For example, P1,4 represents a request for the resource made by node P1 at timestamp 4. The red arrows represent requests, and green arrows represent the corresponding acknowledgment messages. c) [2 points] Suggest a simple method to prevent the algorithm from failing due to message loss.