CS 686 Mid-Semester Exam: Dependable Computing - UVA, Spring 2008 - Prof. John Knight, Exams of Computer Science

A mid-semester examination for the cs 686 dependable computing course at the university of virginia, spring 2008. Instructions for the exam, a pledge, and twelve questions covering various aspects of dependable computing, such as dependability attributes, failure semantics, and fault trees.

Typology: Exams

Pre 2010

Uploaded on 08/30/2009

koofers-user-f0q
koofers-user-f0q 🇺🇸

2.7

(3)

10 documents

1 / 6

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
. This is a closed book, closed notes examination except as noted.
. Write your answers on the examination paper in ink or legible pencil.
.If your answer cannot be read or understood, or if your answer is vague or confused, it will be
marked wrong.
. In any question, make any assumptions that you need to but document your assumptions.
. The number in parentheses after each question is the number of points allocated to that ques-
tion.
NAME (Print Legibly. All Capitals):
PLEDGE (Write Out In Full And Sign):
CS 686
DEPENDABLE COMPUTING
MID-SEMESTER EXAMINATION
SPRING 2008
Total Time Allowed - Two Hours In One Continuous Block
DUE BY 5:00 P.M. ON FRIDAY MARCH 7
pf3
pf4
pf5

Partial preview of the text

Download CS 686 Mid-Semester Exam: Dependable Computing - UVA, Spring 2008 - Prof. John Knight and more Exams Computer Science in PDF only on Docsity!

. This is a closed book, closed notes examination except as noted. . Write your answers on the examination paper in ink or legible pencil. . If your answer cannot be read or understood , or if your answer is vague or confused, it will be marked wrong. . In any question, make any assumptions that you need to but document your assumptions. . The number in parentheses after each question is the number of points allocated to that ques- tion.

NAME ( Print Legibly. All Capitals ):

PLEDGE ( Write Out In Full And Sign ):

CS 686

DEPENDABLE C OMPUTING

MID-S EMESTER EXAMINATION

SPRING 2008

Total Time Allowed - Two Hours In One Continuous Block

D UE BY 5:00 P.M. ON FRIDAY MARCH 7

  1. You have been hired as the dependability engineer for a company that makes information sys- tems for pharmacists. The system provides inventory control of prescription drugs, electronic drug ordering from manufacturers, insurance company billing, and patient records. The sys- tem uses several client computers for pharmacists’ access and a small server that is connected to the Internet. The system supports retail pharmacy operation, and it is used by pharmacy staff constantly. (This is a brief overview. Make assumptions if you need to but state them.)

(a) List the dependability attribute or attributes that need to considered for this system and for each explain why it is important. (5)

(b) For one of the attributes listed in (a), carefully define a possible requirement statement for that dependability attribute for this system. (5)

(c) List the hazard or hazards that are possible for this system. (Remember, a hazard is a state, not an event.) (5)

(d) For one of the hazards, draw the hazard and the top two levels of an example fault tree. (5)

Page score

  1. The taxonomy of definitions by Laprie et al gives the following two definitions, the old and the new , of the critical term dependability:

Old : Dependability is the ability to deliver service that can justifiably be trusted. New : The dependability of a system is the ability to avoid service failures that are more frequent and more severe than is acceptable. Explain why the developers of the taxonomy felt the necessity to make the change from the old to the new. (5)

  1. Carefully and succinctly explain why a mistake in the implementation source code of a piece of software is referred to as a design fault rather than an implementation fault. (5)
  2. Briefly and in your own words, explain what is meant by a Byzantine fault. (5)
  3. What is the “bathtub” curve and how is it useful in developing dependable computer systems? (5)

Page score

  1. A drug infusion pump is used to administer a drug to a patient according to some algorithm that might be demand or schedule driven. What is/are the hazard ( s ) that the system poses? (5)
  2. The notion of failure probability cannot be applied to a software system in the same way that it can be applied to a system subject to degradation faults.

(a) Explain why (5).

(b) Explain how the notion of failure probability is applied to a software system (5).

  1. Compare and contrast the notion of a transient degradation fault with Jim Gray’s notion of a software Heisenbug. (5)