Download CS162 Midterm Exam II - Operating Systems and Systems Programming and more Exams Operating Systems in PDF only on Docsity!
University of California, Berkeley College of Engineering Computer Science Division ⎯ EECS Fall 2006 John Kubiatowicz
Midterm II
December 4th^ , 2006 CS162: Operating Systems and Systems Programming
Your Name:
SID Number: Circle the letters of CS Login
First: a b c d e f g h I j k l m n o p q r s t u v w x y z Second: a b c d e f g h I j k l m n o p q r s t u v w x y z
Discussion Section:
General Information: This is a closed book exam. You are allowed 1 page of hand-written notes (both sides). You have 3 hours to complete as much of the exam as possible. Make sure to read all of the questions first, as some of the questions are substantially more time consuming.
Write all of your answers directly on this paper. Make your answers as concise as possible. On programming questions, we will be looking for performance as well as correctness, so think through your answers carefully. If there is something about the questions that you believe is open to interpretation, please ask us about it!
Problem Possible Score
Total
[ This page left for π ]
Problem 1f[2pts]: “Marshalling” is the process by which Byzantine Generals are forced into making good decisions.
True / False
Explain:
Problem 1g[2pts]: RAID-5 is 5 times more reliable than RAID-1.
True / False
Explain:
Problem 1h[2pts]: One advantage of a software TLB is that the same hardware platform can support both forward and inverse page tables.
True / False
Explain:
Problem 1i[2pts]: The Use bit in the PTE is required for implementing the Second Chance List algorithm.
True / False
Explain:
Problem 1j[2pts]: The difference between a computer Worm and a computer Virus is that a Virus requires human action to spread.
True / False
Explain:
Problem 2: Virtual Memory, Paging, and Disks
Problem 2a[2pts]: Suppose that we have a 64-bit virtual address split as follows:
9 Bits [ Table ID]
13 Bits [ Table ID ]
13 Bits [ Table ID ]
13 Bits [ Page ID ]
16 Bits [ Offset ]
Show the format of a PTE complete with bits required to support the clock algorithm, sharing of dynamic libraries, and copy-on-write optimizations.
Problem 2b[3pts]: Explain how to implement the clock algorithm ( not the Second Chance List algorithm) if the hardware does not directly support use bits or dirty bits.
Problem 2c[4pts]: For the following problem, assume a hypothetical machine with 4 pages of physical memory and 7 pages of virtual memory. Given the access pattern: A B C D E A A E C F F G A C G D C F Indicate in the following table which pages are mapped to which physical pages for each of the following policies. Assume that a blank box matches the element to the left. We have given the FIFO policy as an example.
Access→ A B C D E A A E C F F G A C G D C F 1 A E C 2 B A D 3 C F
FIFO
4 D G
MIN
LRU
Problem 2f[5pts]: Assume that there is a queue in front of the controller. Assume that the operating system behaves as in (2e), and that the distribution of service times has C = 1.5 (i.e. not quite memoryless). Also assume that requests for blocks arrive via an exponential (memoryless) process with an average arrival rate of λ blocks/second. What is the arrival rate λ such that the queue has an average length of 10 blocks? You don’t need to actually compute this number as long as you given an explicit equation for λ. Hint: use Little’s law: Lq =λTq and solve for utilization.
Also, remember from basic algebra that Ax 2 +Bx+C=0 ⇒ x=
A
B B AC
− ±^2 − 4
Problem 3: File Systems Please keep your answers short (one or two sentences per question-mark). We may not give credit for long answers.
Problem 3a[3pts]: Rather than writing updated files to disk immediately when they are closed, many UNIX systems use a delayed write-behind policy in which dirty disk blocks are flushed to disk once every 30 seconds. List two advantages and one disadvantage of such a scheme:
Advantage 1:
Advantage 2:
Disadvantage:
Problem 3b[2pts]: Describe a technique that can be used to mitigate the disadvantage of (3a) without losing the advantages of (3a). Be explicit here.
Problem 3c[4pts]: In the following, compare the performance of a FAT file system to a more primitive system that uses “linked allocation” (where file blocks are linked together via pointers in each block) with respect to number of disk accesses. Be sure to state any assumptions that you are making in terms of what (if anything) is memory resident.
a) Which is faster for random access? Explain. (Assume the file in question is large):
b) Which is faster for sequential access? Explain. (Again assume a large file):
Problem 3h[3pts]: The Andrew File System (AFS) solves the above problem (3g) using state information it maintains at the server. What state is kept? How is that state used to solve the problem?
Problem 3i[4pts]: At time XZ, the request queue for a disk contains the following requests in [track:sector] form:
[10:5], [22:9], [20:21], [21:9], [2:10], [40:45], [6:7], [38:9] …. (in this order).
Assume that the disk head is currently positioned over cylinder 20. What is the sequence of reads under the following head scheduling algorithms?
a) Shortest Seek Time First:
b) SCAN (initially moving upwards):
Problem 4: Security Problem 4a[3pts]: Assume Alice and Bob have never met. Explain how Alice can use a public key infrastructure (PKI) to prove her identity to Bob. Hint: make sure to prevent replay attacks.
Problem 4b[3pts]: Explain how to utilize a PKI to establish a private session key between two parties for fast symmetric encryption.
Problem 4c[4pts]: What are two desirable properties for secure hash functions (ignoring the property where half the bits change for small changes in input). Why are these properties important for signatures?
Problem 5a[4pts]: Under ideal circumstances (and ignoring interrupt and copying overheads and window size), what is the maximum bandwidth that Client A can send data to the server without causing packets to be dropped (Assuming that the headers are of zero length and that routers can handle data at wire speed)? How about Client B? Explain.
Problem 5b[4pts]: Keeping in mind that TCP/IP involves a total header size of 40 bytes (for TCP
- IP), what is the maximum data bandwidth that Client A could send to the server through TCP/IP? Explain. What about Client B? Explain. Again, ignore interrupt and copying overheads. Hint: Assume that a sender runs a protocol that computes the proper maximum packet size to avoid fragmentation within the network.
Problem 5c[6pts]: Now, assume that copying and interrupt overheads impact the maximum rate at which data can be sent or received at the endpoints (clients or servers). Routers continue to be fully pipelined and senders avoid fragmentation as in 5b. What is the maximum data bandwidth that Client A can transfer to the server without dropping packets? Explain. What about Client B? Hint: make sure to account for the impact of the 40 byte TCP/IP header.
Problem 5d[5pts]: Now, accounting for copying and interrupt overheads, what is the total time to reliably send a single maximal-sized packet from Client A to a user-level process on the server without fragmentation? Account for all latencies (including the reception of an ACK, which can be assumed to be zero length but which will generate an interrupt). Explain your work.
Problem 5e[4pts]: Assume the conditions from 5d. Client A sends a continuous stream of packets to the server (and no other clients are talking to the server). How big should the send window be so that the TCP/IP algorithm will achieve maximum bandwidth without dropping packets? Explain. Hint: don’t forget to account for the 40 bytes of header.
