Previous Table of Contents Next


Box 10.2 Reasons for Not Accepting the Results of an Analysis

1.  This needs more analysis.
2.  You need a better understanding of the workload.
3.  It improves performance only for long I/O’s, packets, jobs, and files, and most of the I/O’s, packets, jobs, and files are short.
4.  It improves performance only for short I/O’s, packets, jobs, and files, but who cares for the performance of short I/O’s, packets, jobs, and files; its the long ones that impact the system.
5.  It needs too much memory/CPU/bandwidth and memory/CPU/bandwidth isn’t free.
6.  It only saves us memory/CPU/bandwidth and memory/CPU/bandwidth is cheap.
7.  There is no point in making the networks (similarly, CPUs/disks/...) faster; our CPUs/disks (any component other than the one being discussed) aren’t fast enough to use them.
8.  It improves the performance by a factor of x, but it doesn’t really matter at the user level because everything else is so slow.
9.  It is going to increase the complexity and cost.
10.  Let us keep it simple stupid (and your idea is not stupid).
11.  It is not simple. (Simplicity is in the eyes of the beholder.)
12.  It requires too much state.
13.  Nobody has ever done that before. (You have a new idea.)
14.  It is not going to raise the price of our stock by even an eighth. (Nothing ever does, except rumors.)
15.  This will violate the IEEE, ANSI, CCITT, or ISO standard.
16.  It may violate some future standard.
17.  The standard says nothing about this and so it must not be important.
18.  Our competitors don’t do it. If it was a good idea, they would have done it.
19.  Our competition does it this way and you don’t make money by copying others.
20.  It will introduce randomness into the system and make debugging difficult.
21.  It is too deterministic; it may lead the system into a cycle.
22.  It’s not interoperable.
23.  This impacts hardware.
24.  That’s beyond today’s technology.
25.  It is not self-stabilizing.
26.  Why change—it’s working OK.

EXERCISES

10.1  What type of chart (line or bar) would you use to plot
a.  CPU usage for 12 months of the year
b.  CPU usage as a function of time in months
c.  Number of I/O’s to three disk drives: A, B, and C
d.  Number of I/O’s as a function of number of disk drives in a system
10.2  List the problems with the charts of Figure 10.27.
10.3  A system consists of three resources, called A, B, and C. The measured utilizations are shown in Table 10.3. A zero in a column indicates that the resource is not utilized. Draw a Gantt chart showing utilization profiles.
10.4  The measured values of the eight performance metrics listed in Example 10.2 for a system are 70, 10, 60, 20, 80, 30, 50, and 20%. Draw the Kiviat graph and compute its figure of merit.
10.5  For a computer system of your choice, list a number of HB and LB metrics and draw a typical Kiviat graph using data values of your choice.


FIGURE 10.27  Graphic charts with mistakes.

TABLE 10.3 Data for Exercise 10.3

A B C Time Used(%)

0 0 0 25
0 0 1 10
0 1 0 20
0 1 1 5
1 0 0 5
1 0 1 15
1 1 0 5
1 1 1 15
100


Previous Table of Contents Next

Copyright © John Wiley & Sons, Inc.