1 .\" $NetBSD: methodology.ms,v 1.2 1998/01/09 06:41:00 perry Exp $
3 .\" Copyright (c) 1983 The Regents of the University of California.
4 .\" All rights reserved.
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
9 .\" 1. Redistributions of source code must retain the above copyright
10 .\" notice, this list of conditions and the following disclaimer.
11 .\" 2. Redistributions in binary form must reproduce the above copyright
12 .\" notice, this list of conditions and the following disclaimer in the
13 .\" documentation and/or other materials provided with the distribution.
14 .\" 3. Neither the name of the University nor the names of its contributors
15 .\" may be used to endorse or promote products derived from this software
16 .\" without specific prior written permission.
18 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
19 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
22 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 .\" @(#)methodology.ms 6.2 (Berkeley) 4/16/91
36 Our goal was to evaluate the performance of the target peripherals
37 in an environment as much like our 4.2BSD UNIX systems as possible.
38 There are two basic approaches to creating this kind of test environment.
39 These might be termed the \fIindirect\fR and the \fIdirect\fR approach.
40 The approach used by DEC in producing most of the performance data
41 on the UDA50/RA81 system under VMS is what we term the indirect
43 We chose to use the direct approach.
45 The indirect approach used by DEC involves two steps.
46 First, the environment in which performance is to be evaluated
48 In this case, the disk I/O characteristics of VMS were measured
49 as to the distribution of various sizes of accesses and the proportion
51 This parameterization of
53 I/O activity was termed a
55 The second stage involves simulating this mixture of I/O activities
56 with the devices to be tested and noting the total volume of transactions
57 processed per unit time by each system.
59 The problems encountered with this indirect approach often
60 have to do with the completeness and correctness of the parameterization
61 of the context environment.
64 model constructed for DECs tests uses a random distribution of seeks
65 to the blocks read or written.
66 It is not likely that any real system produces a distribution
67 of disk transfer locations which is truly random and does not
68 exhibit strong locality characteristics.
70 The methodology chosen by us is direct
71 in the sense that it uses the standard structured file system mechanism present
72 in the 4.2BSD UNIX operating system to create the sequence of locations
73 and sizes of reads and writes to the benchmarked equipment.
74 We simply create, write, and read
75 files as they would be by user's activities.
76 The disk space allocation and disk cacheing mechanism built into
77 UNIX is used to produce the actual device reads and writes as well
78 as to determine their size and location on the disk.
79 We measure and compare the rate at which these
83 can be written, rewritten, or read.
85 The advantage of this approach is the implicit accuracy in
86 testing in the same environment in which the peripheral
88 Although this system does not account for the I/O produced
89 by some paging and swapping, in our memory rich environment
90 these activities account for a relatively small portion
91 of the total disk activity.
93 A more significant disadvantage to the direct approach
94 is the occasional difficulty we have in accounting for our
96 The apparently straight-forward activity of reading or writing a logical file
97 on disk can produce a complex mixture of disk traffic.
98 File I/O is supported by a file management system that
99 buffers disk traffic through an internal cache,
100 which allows writes to ba handled asynchronously.
101 Reads must be done synchronously,
102 however this restriction is moderated by the use of read-ahead.
103 Small changes in the performance of the disk controller
104 subsystem can result in large and unexpected
105 changes in the file system performance,
106 as it may change the characteristics of the memory contention
107 experienced by the processor.