You have discovered an interesting new organism while examining the surface of a 2×4 piece of wood. Which of the following criteria will allow you to classify the organism as belonging to Bacteria but not Archaea or Eukarya?

You have discovered a new prokaryote. How would you identify the organism as a cyanobacterium?

Potpourri_2a Potpourri 2. [8 points] Answer the following questions with reference to Xen’s I/O ring data structure. The I/O ring is a circular buffer used by Guest Operating systems for enqueuing requests for the Hypervisor, and collecting its responses. a. [2 points] You wish to improve I/O ring design by providing APIs in the guest OS to throttle the request creation. Assume the Guest OS needs just the number of outstanding requests to decide if it should throttle the requests. Is it possible to get this number of outstanding requests in the current implementation? Why or why not?

Parallel_Systems_3b M.E.Lock The context for this question is the same as the previous question: [4 points] The Linux kernel historically used a ticket spin lock on multiprocessor systems. In 2016, Linux patch introduced a “qspinlock”, which is a “hybrid” design drawing inspiration from queue-based locks (MCS) and the old ticket lock.  “qspinlock” is built based on a linked-list lock. However, the first spinner (next-in-line thread) spins on a lock bit instead of carrying a node structure from lock to unlock. When more contenting thread arrives, the first spinner pays the overhead to revert back to spin on the node structure and queues additional threads in a linked list. b. [2 point] Describe a scenario when either a ticket lock or a pure queue-based lock may outperform qspinlock. Explain why. 

Parallel_Systems_6a LRPC 6. [6 points] In the given figure, FCFS stands for First Come First Serve, FP stands for Fixed Processor, and LP stands for Last Processor.  a. [2 points] Compare and contrast First Come First Served (FCFS) and Fixed Processor (FP) scheduling approaches on the metrics of cache affinity and load-balancing. (succinct bullets please) 

OS_Structure_4b Microkernel The context for this question is same as the previous question. 4. You are building an OS using a microkernel-based approach following the principles of the L3 microkernel. The processor architecture you are building this OS for has the following features:      • A 32-bit hardware address space.    • Paged virtual memory system (8KB pages) with a processor register called PTBR that points to the page table in memory to enable hardware address translation.     • A TLB with Address space IDs associated with each TLB entry.     • A pair of hardware-enforced segment registers (lower and upper bound of virtual addresses) which limit the virtual address space that can be accessed by a process running on the processor.     • A virtually-indexed physically tagged processor cache. You end up with 2 big subsystems (A and B) that each require 230 bytes of virtual memory space. You also end up with 4 subsystems (C,D,E,F) that require 100×220 , 500×220 , 1000×220 , and 2000×220 bytes of virtual memory, respectively. You want to put each of these subsystems in their own protection domains.  b. [2 points] (Answer this question based on your grouping of the protection domains) Suppose there is a context switch from A to C. What does your OS do to facilitate this context switch? 

Parallel_Systems_1a Shared Memory Machines 1. [6 points] a. [4 points] You are trying to explain the principles of sequential consistency and cache coherence to your friend. Your friend claims that cache coherence implies sequential consistency. Do you agree with your friend’s claim? If not, state how you will convince your friend that they are not the same. 

Parallel_Systems_4 Barriers 4. [3 points] Consider the sense reversing barrier implementation in C as follows where count, sense are shared variables which are initialized to N and False respectively. Will this code work? If yes explain why, if no, give brief reasoning and an example! 

Virtualization_2b Full Virtualization   The context for this question is the same as the previous question: 2. CPU virtualization mechanisms in a virtualized environment have two primary objectives. Providing each guest OS with the illusion of exclusive CPU ownership.  Delivering program discontinuity events to the guest OS.   b. [1 point] How are these events delivered to the guest OS in a fully virtualized setting?

There are many approaches to addressing the issue of bullying. Research indicates that one effective approach is: