Previous Table of Contents Next


TABLE 19.10 Measured Throughputs for Scheduler Design Study

Experiment No. A B C D E TW TI TB

1 –1 –1 –1 –1 1 15.0 25.0 15.2
2 1 –1 –1 –1 –1 11.0 41.0 3.0
3 –1 1 –1 –1 –1 25.0 36.0 21.0
4 1 1 –1 –1 1 10.0 15.7 8.6
5 –1 –1 1 –1 –1 14.0 63.9 7.5
6 1 –1 1 –1 1 10.0 13.2 7.5
7 –1 1 1 –1 1 28.0 36.3 20.2
8 1 1 1 –1 –1 11.0 23.0 3.0
9 –1 –1 –1 1 –1 14.0 66.1 6.4
10 1 –1 –1 1 1 10.0 9.1 8.4
11 –1 1 –1 1 1 27.0 34.6 15.7
12 1 1 –1 1 –1 11.0 23.0 3.0
13 –1 –1 1 1 1 14.0 26.0 12.0
14 1 –1 1 1 –1 11.0 38.0 2.0
15 –1 1 1 1 –1 25.0 35.0 17.2
16 1 1 1 1 1 11.0 22.0 2.0

TABLE 19.11 Effects and Variation Explained In Scheduling Policy Comparison Study

Confounded
Effects
TW
TI
TB
1 2 Estimate % of Variation Estimate % of Variation Estimate % of Variation

I ABCDE 15.44 31.74 9.54
A BCDE –4.81 55.5 –8.62 31.0 –4.86 58.8
B ACDE 3.06 22.5 –3.54 5.2 1.79 8.0
C ABDE 0.06 0.0 0.43 0.1 –0.62 1.0
D ABCE –0.06 0.0 –0.02 0.0 –1.21 3.6
AB CDE –2.94 20.7 1.34 0.8 –2.33 13.5
AC BDE 0.06 0.0 0.49 0.1 –0.44 0.5
AD BCE 0.19 0.1 –0.08 0.0 0.37 0.3
BC ADE 0.19 0.1 0.44 0.1 –0.12 0.0
BD ACE 0.06 0.0 0.47 0.1 –0.66 1.1
CD ABE –0.19 0.1 –1.91 1.5 0.58 0.8
DE ABC –0.06 0.0 0.21 0.0 –0.47 0.5
CE ABD 0.06 0.0 1.21 0.6 –0.16 0.1
BE ACD 0.31 0.2 7.96 26.4 –1.37 4.7
AE BCD –0.56 0.8 0.88 0.3 0.28 0.2
E ABCD 0.19 0.1 –9.01 33.8 1.66 6.8

  Factor C (Queue assignment) or any of its interactions do not have any significant impact on the throughput.
  Factor D (Requeuing) also is not very effective.
  Factor A (Preemption) impacts all three workloads significantly.
  Factor B (Time slice) has less impact than factor A.
  Factor E (Fairness) is important for interactive jobs and slightly important for background jobs.

We will reconsider this case study in Section 23.3, where we present further conclusions.

EXERCISES

19.1  Analyze the 24–1 design shown in Table 19.12.
TABLE 19.12 A 24–1 Design

C1D1 C1D2 C2D1 C2D2

A1B1 40 15
A1B2 20 10
A2B1 100 30
A2B2 120 50

a.  Quantify all main effects.
b.  Quantify percentages of variation explained.
c.  Sort the variables in the order of decreasing importance.
d.  List all confoundings.
e.  Can you propose a better design with the same number of experiments.
f.  What is the resolution of the design?
19.2  Is it possible to have a 2III4–1 design? A 2II4–1 design? A 2IV4–1 design? If yes, give an example.


Previous Table of Contents Next

Copyright © John Wiley & Sons, Inc.