Home Explore Help
Sign In
zorun
/
kea
1
0
Fork 0
Files Issues 0 Pull Requests 0 Wiki
Tree: e7c00e175f
Branches Tags
kea
/
tests
/
tools
/
dhcp-ubench
/
dhcp-perf-guide.html

dhcp-perf-guide.html 50 KB
History Raw

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624 
  1. <html><head>
    2. 

  2.       <meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
    3. 

  3.    <title>DHCP Performance Guide</title><link rel="stylesheet" type="text/css" href="bind10-guide.css"><meta name="generator" content="DocBook XSL Stylesheets V1.76.1"><meta name="description" content="BIND 10 is a framework that features Domain Name System (DNS) suite and Dynamic Host Configuration Protocol (DHCP) servers with development managed by Internet Systems Consortium (ISC). This document describes various aspects of DHCP performance, measurements and tuning. It covers BIND 10 DHCP (codename Kea), existing ISC DHCP4 software, perfdhcp (a DHCP performance measurement tool) and other related topics."></head><body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF"><div class="book" title="DHCP Performance Guide"><div class="titlepage"><div><div><h1 class="title"><a name="d0e3"></a>DHCP Performance Guide</h1></div><div><div class="author"><h3 class="author"><span class="firstname">Tomasz</span> <span class="surname">Mrugalski</span></h3></div></div><div><div class="author"><h3 class="author"><span class="firstname">Marcin</span> <span class="surname">Siodelski</span></h3></div></div><div><p class="releaseinfo">This is a companion document for BIND 10 version
    4. 

  4.     20120817.</p></div><div><p class="copyright">Copyright &copy; 2012 Internet Systems Consortium, Inc. ("ISC")</p></div><div><div class="abstract" title="Abstract"><p class="title"><b>Abstract</b></p><p>BIND 10 is a framework that features Domain Name System
    5. 

  5.       (DNS) suite and Dynamic Host Configuration Protocol (DHCP)
    6. 

  6.       servers with development managed by Internet Systems Consortium (ISC).
    7. 

  7.       This document describes various aspects of DHCP performance,
    8. 

  8.       measurements and tuning. It covers BIND 10 DHCP (codename Kea),
    9. 

  9.       existing ISC DHCP4 software, perfdhcp (a DHCP performance
    10. 

  10.       measurement tool) and other related topics.</p></div></div></div><hr></div><div class="toc"><p><b>Table of Contents</b></p><dl><dt><span class="preface"><a href="#d0e28">Preface</a></span></dt><dd><dl><dt><span class="section"><a href="#acknowledgements">Acknowledgements</a></span></dt></dl></dd><dt><span class="chapter"><a href="#intro">1. Introduction</a></span></dt><dt><span class="chapter"><a href="#dhcp4">2. ISC DHCP 4.x</a></span></dt><dt><span class="chapter"><a href="#kea">3. Kea</a></span></dt><dd><dl><dt><span class="section"><a href="#d0e54">Backend performance evaluation</a></span></dt><dt><span class="section"><a href="#mysql-backend">MySQL backend</a></span></dt><dd><dl><dt><span class="section"><a href="#d0e156">MySQL tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#sqlite-ubench">SQLite-ubench</a></span></dt><dd><dl><dt><span class="section"><a href="#sqlite-tweaks">SQLite tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#memfile-ubench">memfile-ubench</a></span></dt><dd><dl><dt><span class="section"><a href="#memfile-tweaks">memfile tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#d0e245">Performance measurements</a></span></dt><dt><span class="section"><a href="#d0e521">Possible further optimizations</a></span></dt></dl></dd><dt><span class="chapter"><a href="#perfdhcp">4. perfdhcp</a></span></dt><dd><dl><dt><span class="section"><a href="#d0e533">Purpose</a></span></dt><dt><span class="section"><a href="#perfdhcp-key-features">Key features</a></span></dt><dt><span class="section"><a href="#perfdhcp-command-line">Command line options</a></span></dt><dt><span class="section"><a href="#starting-perfdhcp">Starting perfdhcp</a></span></dt><dt><span class="section"><a href="#perfdhcp-commandline-examples">perfdhcp command line examples</a></span></dt><dd><dl><dt><span class="section"><a href="#perfdhcp-basic-usage">Example: basic usage</a></span></dt><dt><span class="section"><a href="#perfdhcp-rate-control">Example: rate control</a></span></dt><dt><span class="section"><a href="#perfdhcp-templates">Example: templates</a></span></dt></dl></dd></dl></dd></dl></div><div class="list-of-tables"><p><b>List of Tables</b></p><dl><dt>3.1. <a href="#d0e256">Synchronous results</a></dt><dt>3.2. <a href="#d0e332">Asynchronous results</a></dt><dt>3.3. <a href="#tbl-perf-results">Estimated performance</a></dt></dl></div><div class="preface" title="Preface"><div class="titlepage"><div><div><h2 class="title"><a name="d0e28"></a>Preface</h2></div></div></div><div class="toc"><p><b>Table of Contents</b></p><dl><dt><span class="section"><a href="#acknowledgements">Acknowledgements</a></span></dt></dl></div><div class="section" title="Acknowledgements"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="acknowledgements"></a>Acknowledgements</h2></div></div></div><p>ISC would like to acknowledge generous support for
    11. 

  11.       BIND 10 development of DHCPv4 and DHCPv6 components provided
    12. 

  12.       by <a class="ulink" href="http://www.comcast.com/" target="_top">Comcast</a>.</p></div></div><div class="chapter" title="Chapter&nbsp;1.&nbsp;Introduction"><div class="titlepage"><div><div><h2 class="title"><a name="intro"></a>Chapter&nbsp;1.&nbsp;Introduction</h2></div></div></div><p>
    13. 

  13.       This document is in its early stages of development. It is
    14. 

  14.       expected to grow significantly in a near future. It will
    15. 

  15.       cover topics like database backend perfomance measurements,
    16. 

  16.       pros an cons of various optimization techniques and
    17. 

  17.       tools.
    18. 

  18.     </p></div><div class="chapter" title="Chapter&nbsp;2.&nbsp;ISC DHCP 4.x"><div class="titlepage"><div><div><h2 class="title"><a name="dhcp4"></a>Chapter&nbsp;2.&nbsp;ISC DHCP 4.x</h2></div></div></div><p>
    19. 

  19.       TODO: Write something about ISC DHCP4 here.
    20. 

  20.     </p></div><div class="chapter" title="Chapter&nbsp;3.&nbsp;Kea"><div class="titlepage"><div><div><h2 class="title"><a name="kea"></a>Chapter&nbsp;3.&nbsp;Kea</h2></div></div></div><div class="toc"><p><b>Table of Contents</b></p><dl><dt><span class="section"><a href="#d0e54">Backend performance evaluation</a></span></dt><dt><span class="section"><a href="#mysql-backend">MySQL backend</a></span></dt><dd><dl><dt><span class="section"><a href="#d0e156">MySQL tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#sqlite-ubench">SQLite-ubench</a></span></dt><dd><dl><dt><span class="section"><a href="#sqlite-tweaks">SQLite tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#memfile-ubench">memfile-ubench</a></span></dt><dd><dl><dt><span class="section"><a href="#memfile-tweaks">memfile tweaks</a></span></dt></dl></dd><dt><span class="section"><a href="#d0e245">Performance measurements</a></span></dt><dt><span class="section"><a href="#d0e521">Possible further optimizations</a></span></dt></dl></div><p>
    21. 

  21. 

  22.     </p><div class="section" title="Backend performance evaluation"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="d0e54"></a>Backend performance evaluation</h2></div></div></div><p>
    23. 

  23.         Kea will support several different database backends, using
    24. 

  24.         both popular databases (like MySQL or SQLite) and
    25. 

  25.         custom-developed solutions (like in-memory database).  BIND 10
    26. 

  26.         source code features set of performance microbenchmarks.
    27. 

  27.         These are small tools written in C/C++ that simulate expected
    28. 

  28.         DHCP server behaviour and evaluate the performance of
    29. 

  29.         considered databases. As implemented benchmarks are not really
    30. 

  30.         simulating DHCP operation, but rather use set of primitives
    31. 

  31.         that can be used by a real server, they are called
    32. 

  32.         micro-benchmarks.
    33. 

  33.       </p><p>Although there are many operations and data types that
    34. 

  34.       server could store in a database, the most frequently used data
    35. 

  35.       type is lease information. Although lease information for IPv4
    36. 

  36.       and IPv6 differs slightly, it is expected that the performance
    37. 

  37.       differences will be minimal between IPv4 and IPv6 lease operations.
    38. 

  38.       Therefore each test uses lease4 table for performance measurements.
    39. 

  39.       </p><p>All benchmarks are implemented as single threaded applications
    40. 

  40.       that take advantage of a single database connection.</p><p>
    41. 

  41.         Those benchmarks are stored in tests/tools/dhcp-ubench
    42. 

  42.         directory. This directory contains simplified prototypes for
    43. 

  43.         various DB back-ends that are planned or considered as a
    44. 

  44.         backend engine for BIND10 DHCP.  Athough trivial now, they are
    45. 

  45.         expected to evolve into useful tools that will allow users to
    46. 

  46.         measure performance in their specific environment.
    47. 

  47.       </p><p>
    48. 

  48.       Currently the following benchmarks are implemented:
    49. 

  49.       </p><div class="itemizedlist"><ul class="itemizedlist" type="disc"><li class="listitem"><p>in memory+flat file</p></li><li class="listitem"><p>SQLite</p></li><li class="listitem"><p>MySQL</p></li></ul></div><p>
    50. 

  50.     </p><p>
    51. 

  51.       As they require additional (sometimes heavy) dependencies, they are not
    52. 

  52.       built by default. Actually, their build system is completely separated.
    53. 

  53.       It will be eventually merged with the main BIND10 makefile system, but
    54. 

  54.       that is a low priority for now.
    55. 

  55.     </p><p>
    56. 

  56.       All benchmarks will follow the same pattern:
    57. 

  57.       </p><div class="orderedlist"><ol class="orderedlist" type="1"><li class="listitem"><p>prepare operation (connect to a database, create a file etc.)</p></li><li class="listitem"><p>Measure timestamp 0</p></li><li class="listitem"><p>Commit new lease4 (repeated X times)</p></li><li class="listitem"><p>Measure timestamp 1</p></li><li class="listitem"><p>Search for random lease4 (repeated X times)</p></li><li class="listitem"><p>Measure timestamp 2</p></li><li class="listitem"><p>Update existing lease4 (repeated X times)</p></li><li class="listitem"><p>Measure timestamp 3</p></li><li class="listitem"><p>Delete existing lease4 (repeated X times)</p></li><li class="listitem"><p>Measure timestamp 4</p></li><li class="listitem"><p>Print out statistics, based on X and measured timestamps.</p></li></ol></div><p>
    58. 

  58. 

  59.       Although this approach does not attempt to simulate actual DHCP server
    60. 

  60.       operation that has mix of all steps intervening, it answers the
    61. 

  61.       questions about basic database strenghts and weak points. In particular
    62. 

  62.       it can show what is the impact of specific DB optimizations, like
    63. 

  63.       changing engine, optimizing for writes/reads etc.
    64. 

  64.     </p><p>
    65. 

  65.       The framework attempts to do the same amount of operations for every
    66. 

  66.       backend thus allowing fair complarison between them.
    67. 

  67.     </p></div><div class="section" title="MySQL backend"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="mysql-backend"></a>MySQL backend</h2></div></div></div><p>MySQL backend requires MySQL client development libraries. It uses
    68. 

  68.       mysql_config tool (that works similar to pkg-config) to discover required
    69. 

  69.       compilation and linking options. To install required packages on Ubuntu,
    70. 

  70.       use the following command:
    71. 

  71. 

  72.       </p><pre class="screen">$ <strong class="userinput"><code>sudo apt-get install mysql-client mysql-server libmysqlclient-dev</code></strong></pre><p>
    73. 

  73. 

  74.       Make sure that MySQL server is running. Make sure that you have your setup
    75. 

  75.       configured so there is a user that is able to modify used database.</p><p>Before running tests, you need to initialize your database. You can
    76. 

  76.       use mysql.schema script for that purpose. WARNING: It will drop existing
    77. 

  77.       Kea database. Do not run this on your production server. Assuming your
    78. 

  78.       MySQL user is kea, you can initialize your test database by:
    79. 

  79. 

  80.       </p><pre class="screen">$ <strong class="userinput"><code>mysql -u kea -p &lt; mysql.schema</code></strong></pre><p>
    81. 

  81.       </p><p>After database is initialized, you are ready to run the test:
    82. 

  82.       </p><pre class="screen">$ <strong class="userinput"><code>./mysql_ubench</code></strong></pre><p>
    83. 

  83. 

  84.       or
    85. 

  85. 

  86.       </p><pre class="screen">$ <strong class="userinput"><code>./mysql_ubench &gt; results-&gt;mysql.txt</code></strong></pre><p>
    87. 

  87. 

  88.       Redirecting output to a file is important, because for each operation
    89. 

  89.       there is a single character printed to show progress. If you have a slow
    90. 

  90.       terminal, this may considerably affect test perfromance. On the other hand,
    91. 

  91.       printing something after each operation is required, as poor DB setting
    92. 

  92.       may slow down operations to around 20 per second. Observant user is expected
    93. 

  93.       to note that initial dots are printed too slowly and abort the test.</p><p>Currently all default parameters are hardcoded. Default values can be
    94. 

  94.       overwritten using command line switches. Although all benchmarks take
    95. 

  95.       the same list of parameters, some of them are specific to a given backend
    96. 

  96.       type. To get a list of supported parameters, run your benchmark with -h option:
    97. 

  97. 

  98.       </p><pre class="screen">$ <strong class="userinput"><code>./mysql_ubench -h</code></strong>
    99. 

  99. This is a benchmark designed to measure expected performance
    100. 

  100. of several backends. This particular version identifies itself
    101. 

  101. as following:
    102. 

  102. MySQL client version is 5.5.24
    103. 

  103. 

  104. Possible command-line parameters:
    105. 

  105.  -h - help (you are reading this)
    106. 

  106.  -m hostname - specifies MySQL server to connect (MySQL backend only)
    107. 

  107.  -u username - specifies MySQL user name (MySQL backend only)
    108. 

  108.  -p password - specifies MySQL passwod (MySQL backend only)
    109. 

  109.  -f name - database or filename (MySQL, SQLite and memfile)
    110. 

  110.  -n integer - number of test repetitions (MySQL, SQLite and memfile)
    111. 

  111.  -s yes|no - synchronous/asynchronous operation (MySQL, SQLite and memfile)
    112. 

  112.  -v yes|no - verbose mode (MySQL, SQLite and memfile)
    113. 

  113.  -c yes|no - should compiled statements be used (MySQL only)
    114. 

  114. </pre><p>
    115. 

  115. 

  116.       </p><div class="section" title="MySQL tweaks"><div class="titlepage"><div><div><h3 class="title"><a name="d0e156"></a>MySQL tweaks</h3></div></div></div><p>One parameter that has huge impact on performance is a a backend engine.
    117. 

  117.         You can get a list of engines of your MySQL implementation by using
    118. 

  118. 

  119.         </p><pre class="screen">&gt; <strong class="userinput"><code>show engines;</code></strong></pre><p>
    120. 

  120. 

  121.         in your mysql client. Two notable engines are MyISAM and InnoDB. mysql_ubench will
    122. 

  122.         use MyISAM for synchronous mode and InnoDB for asynchronous.</p></div></div><div class="section" title="SQLite-ubench"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="sqlite-ubench"></a>SQLite-ubench</h2></div></div></div><p>SQLite backend requires both sqlite3 development and run-time package. Their
    123. 

  123.       names may vary from system to system, but on Ubuntu 12.04 they are called
    124. 

  124.       sqlite3 libsqlite3-dev. To install them, use the following command:
    125. 

  125. 

  126.       </p><pre class="screen">&gt; <strong class="userinput"><code>sudo apt-get install sqlite3 libsqlite3-dev</code></strong></pre><p>
    127. 

  127. 

  128.       Before running the test the database has to be created. Use the following command for that:
    129. 

  129.       </p><pre class="screen">&gt; <strong class="userinput"><code>cat sqlite.schema | sqlite3 sqlite.db</code></strong></pre><p>
    130. 

  130. 

  131.       A new database called sqlite.db will be created. That is the default name used
    132. 

  132.       by sqlite_ubench test. If you prefer other name, make sure you update
    133. 

  133.       sqlite_ubench.cc accordingly.</p><p>Once the database is created, you can run tests:
    134. 

  134.       </p><pre class="screen">&gt; <strong class="userinput"><code>./sqlite_ubench</code></strong></pre><p>
    135. 

  135.       or
    136. 

  136.       </p><pre class="screen">&gt; <strong class="userinput"><code>./sqlite_ubench &gt; results-sqlite.txt</code></strong></pre><p>
    137. 

  137.       </p><div class="section" title="SQLite tweaks"><div class="titlepage"><div><div><h3 class="title"><a name="sqlite-tweaks"></a>SQLite tweaks</h3></div></div></div><p>To modify default sqlite_ubench parameters, command line
    138. 

  138.         switches can be used. Currently supported parameters are
    139. 

  139.         (default values specified in brackets):
    140. 

  140.         </p><div class="orderedlist"><ol class="orderedlist" type="1"><li class="listitem"><p>-f filename - name of the database file ("sqlite.db")</p></li><li class="listitem"><p>-n num - number of iterations (100)</p></li><li class="listitem"><p>-s yes|no - should the operations be performend in synchronous (yes)
    141. 

  141.           or asynchronous (no) manner (yes)</p></li><li class="listitem"><p>-v yes|no - verbose mode. Should the test print out progress? (yes)</p></li><li class="listitem"><p>-c yes|no - compiled statements. Should the SQL statements be precompiled?</p></li></ol></div><p>
    142. 

  142.         </p><p>SQLite can run in asynchronous or synchronous mode. This
    143. 

  143.         mode can be controlled by using sync parameter. It is set
    144. 

  144.         using (PRAGMA synchronous = ON or OFF).</p><p>Another tweakable feature is journal mode. It can be
    145. 

  145.         turned to several modes of operation. Its value can be
    146. 

  146.         modified in SQLite_uBenchmark::connect().  See
    147. 

  147.         http://www.sqlite.org/pragma.html#pragma_journal_mode for
    148. 

  148.         detailed explanantion.</p></div></div><div class="section" title="memfile-ubench"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="memfile-ubench"></a>memfile-ubench</h2></div></div></div><p>Memfile backend is custom developed prototype backend that
    149. 

  149.       somewhat mimics operation of ISC DHCP4. It uses in-memory
    150. 

  150.       storage using standard C++ and boost mechanisms (std::map and
    151. 

  151.       boost::shared_ptr&lt;&gt;). All database changes are also
    152. 

  152.       written to a lease file. That file is strictly write-only. This
    153. 

  153.       approach takes advantage of the fact that simple append is faster
    154. 

  154.       than edition with potential whole file relocation.</p><div class="section" title="memfile tweaks"><div class="titlepage"><div><div><h3 class="title"><a name="memfile-tweaks"></a>memfile tweaks</h3></div></div></div><p>To modify default memfile_ubench parameters, command line
    155. 

  155.         switches can be used. Currently supported parameters are
    156. 

  156.         (default values specified in brackets):
    157. 

  157.         </p><div class="orderedlist"><ol class="orderedlist" type="1"><li class="listitem"><p>-f filename - name of the database file ("dhcpd.leases")</p></li><li class="listitem"><p>-n num - number of iterations (100)</p></li><li class="listitem"><p>-s yes|no - should the operations be performend in synchronous (yes)
    158. 

  158.           or asynchronous (no) manner (yes)</p></li><li class="listitem"><p>-v yes|no - verbose mode. Should the test print out progress? (yes)</p></li></ol></div><p>
    159. 

  159.         </p><p>memfile can run in asynchronous or synchronous mode. This
    160. 

  160.         mode can be controlled by using sync parameter. It uses
    161. 

  161.         fflush() and fsync() in synchronous mode to make sure that
    162. 

  162.         data is not buffered and physically stored on disk.</p></div></div><div class="section" title="Performance measurements"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="d0e245"></a>Performance measurements</h2></div></div></div><p>This section contains sample results for backend performance measurements,
    163. 

  163.       taken using microbenchmarks. Tests were conducted on reasonably powerful machine:
    164. 

  164.       </p><pre class="screen">
    165. 

  165. CPU: Quad-core Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz (8 logical cores)
    166. 

  166. HDD: 1,5TB Seagate Barracuda ST31500341AS 7200rpm (used only one of them), ext4 partition
    167. 

  167. OS: Ubuntu 12.04, running kernel 3.2.0-26-generic SMP x86_64
    168. 

  168. compiler: g++ (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3
    169. 

  169. MySQL version: 5.5.24
    170. 

  170. SQLite version: 3.7.9sourceid version is 2011-11-01 00:52:41 c7c6050ef060877ebe77b41d959e9df13f8c9b5e</pre><p>
    171. 

  171.       </p><p>Benchmarks were run in two series: synchronous and
    172. 

  172.       asynchronous. As those modes offer radically different
    173. 

  173.       performances, synchronous mode was conducted for 1000 (one
    174. 

  174.       thousand) repetitions and asynchronous mode was conducted for
    175. 

  175.       100000 (hundred thousand) repetitions.</p><div class="table"><a name="d0e256"></a><p class="title"><b>Table&nbsp;3.1.&nbsp;Synchronous results</b></p><div class="table-contents"><table summary="Synchronous results" border="1"><colgroup><col align="center" class="Backend"><col align="center" class="Num"><col align="center" class="Create"><col align="center" class="Search"><col align="center" class="Update"><col align="center" class="Delete"></colgroup><thead><tr><th align="center">Backend</th><th align="center">Operations</th><th align="center">Create</th><th align="center">Search</th><th align="center">Update</th><th align="center">Delete</th><th align="center">Average</th></tr></thead><tbody><tr><td align="center">MySQL</td><td align="center">1000</td><td align="center">31.603978s</td><td align="center"> 0.116612s</td><td align="center">27.964191s</td><td align="center">27.695209s</td><td align="center">21.844998s</td></tr><tr><td align="center">SQLite</td><td align="center">1000</td><td align="center">61.421356s</td><td align="center"> 0.033283s</td><td align="center">59.476638s</td><td align="center">56.034150s</td><td align="center">44.241357s</td></tr><tr><td align="center">memfile</td><td align="center">1000</td><td align="center">41.711886s</td><td align="center"> 0.000724s</td><td align="center">42.267578s</td><td align="center">42.169679s</td><td align="center">31.537467s</td></tr></tbody></table></div></div><br class="table-break"><p>Following parameters were measured for asynchronous mode.
    176. 

  176.       MySQL and SQLite were run with 100 thousand repetitions. Memfile
    177. 

  177.       was run for 1 million repetitions due to much larger performance.</p><div class="table"><a name="d0e332"></a><p class="title"><b>Table&nbsp;3.2.&nbsp;Asynchronous results</b></p><div class="table-contents"><table summary="Asynchronous results" border="1"><colgroup><col align="center" class="Backend"><col align="center" class="Num"><col align="center" class="Create"><col align="center" class="Search"><col align="center" class="Update"><col align="center" class="Delete"></colgroup><thead><tr><th align="center">Backend</th><th align="center">Operations</th><th align="center">Create [s]</th><th align="center">Search [s]</th><th align="center">Update [s]</th><th align="center">Delete [s]</th><th align="center">Average [s]</th></tr></thead><tbody><tr><td align="center">MySQL</td><td align="center">100000</td><td align="center">10.584842s</td><td align="center">10.386402s</td><td align="center">10.062384s</td><td align="center"> 8.890197s</td><td align="center"> 9.980956s</td></tr><tr><td align="center">SQLite</td><td align="center">100000</td><td align="center"> 3.710356s</td><td align="center"> 3.159129s</td><td align="center"> 2.865354s</td><td align="center"> 2.439406s</td><td align="center"> 3.043561s</td></tr><tr><td align="center">memfile</td><td align="center">1000000 (sic!)</td><td align="center"> 6.084131s</td><td align="center"> 0.862667s</td><td align="center"> 6.018585s</td><td align="center"> 5.146704s</td><td align="center"> 4.528022s</td></tr></tbody></table></div></div><br class="table-break"><p>Presented performance results can be computed into operations per second metrics.
    178. 

  178.       It should be noted that due to large differences between various operations (sometime
    179. 

  179.       over 3 orders of magnitude), it is difficult to create a simple, readable chart with
    180. 

  180.       that data.</p><div class="table"><a name="tbl-perf-results"></a><p class="title"><b>Table&nbsp;3.3.&nbsp;Estimated performance</b></p><div class="table-contents"><table summary="Estimated performance" border="1"><colgroup><col align="center" class="Backend"><col align="center" class="Create"><col align="center" class="Search"><col align="center" class="Update"><col align="center" class="Delete"><col align="center" class="Average"></colgroup><thead><tr><th align="center">Backend</th><th align="center">Create [oper/s]</th><th align="center">Search [oper/s]</th><th align="center">Update [oper/s]</th><th align="center">Delete [oper/s]</th><th align="center">Average [oper/s]</th></tr></thead><tbody><tr><td align="center">MySQL (async)</td><td align="center">9447.47</td><td align="center">9627.97</td><td align="center">9938.00</td><td align="center">11248.34</td><td align="center">10065.45</td></tr><tr><td align="center">SQLite (async)</td><td align="center">26951.59</td><td align="center">31654.29</td><td align="center">34899.70</td><td align="center">40993.59</td><td align="center">33624.79</td></tr><tr><td align="center">memfile (async)</td><td align="center">164362.01</td><td align="center">1159195.84</td><td align="center">166152.01</td><td align="center">194299.11</td><td align="center">421002.24</td></tr><tr><td align="center">MySQL (sync)</td><td align="center">31.64</td><td align="center">8575.45</td><td align="center">35.76</td><td align="center">36.11</td><td align="center">2169.74</td></tr><tr><td align="center">SQLite (sync)</td><td align="center">16.28</td><td align="center">20045.37</td><td align="center">16.81</td><td align="center">17.85</td><td align="center">7524.08</td></tr><tr><td align="center">memfile (sync)</td><td align="center">23.97</td><td align="center">1381215.47</td><td align="center">23.66</td><td align="center">23.71</td><td align="center">345321.70</td></tr></tbody></table></div></div><br class="table-break"><div class="mediaobject"><img src="performance-results-graph1.png" alt="Performance measurements"><div class="caption"><p>Graphical representation of the performance results
    181. 

  181.           presented in table <a class="xref" href="#tbl-perf-results" title="Table&nbsp;3.3.&nbsp;Estimated performance">Table&nbsp;3.3, &#8220;Estimated performance&#8221;</a>.</p></div></div></div><div class="section" title="Possible further optimizations"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="d0e521"></a>Possible further optimizations</h2></div></div></div><p>
    182. 

  182.         For debugging purposes the code was compiled with -g -O0
    183. 

  183.         flags. While majority of the time was spent in backend
    184. 

  184.         functions (that was probably compiled with -O2 flags), the
    185. 

  185.         benchmark code could perform faster, when compiled with -O2,
    186. 

  186.         rather than -O0. That is expected to affect memfile benchmark.
    187. 

  187.       </p><p>
    188. 

  188.         Currently all operations were conducted on one by one
    189. 

  189.         basis. Each operation was treated as a separate
    190. 

  190.         transaction. Grouping X operations together will potentially
    191. 

  191.         bring almost X fold increase in synchronous operations.
    192. 

  192.         Extension for this benchmark in this regard should be considered.
    193. 

  193.         That affects only write operations (insert, update and delete). Read
    194. 

  194.         operations (search) are expected to be barely affected.
    195. 

  195.       </p><p>
    196. 

  196.         Multi-threaded or multi-process benchmark may be considered in
    197. 

  197.         the future. It may be somewhat difficult as only some backends
    198. 

  198.         support concurrent access.
    199. 

  199.       </p></div></div><div class="chapter" title="Chapter&nbsp;4.&nbsp;perfdhcp"><div class="titlepage"><div><div><h2 class="title"><a name="perfdhcp"></a>Chapter&nbsp;4.&nbsp;perfdhcp</h2></div></div></div><div class="toc"><p><b>Table of Contents</b></p><dl><dt><span class="section"><a href="#d0e533">Purpose</a></span></dt><dt><span class="section"><a href="#perfdhcp-key-features">Key features</a></span></dt><dt><span class="section"><a href="#perfdhcp-command-line">Command line options</a></span></dt><dt><span class="section"><a href="#starting-perfdhcp">Starting perfdhcp</a></span></dt><dt><span class="section"><a href="#perfdhcp-commandline-examples">perfdhcp command line examples</a></span></dt><dd><dl><dt><span class="section"><a href="#perfdhcp-basic-usage">Example: basic usage</a></span></dt><dt><span class="section"><a href="#perfdhcp-rate-control">Example: rate control</a></span></dt><dt><span class="section"><a href="#perfdhcp-templates">Example: templates</a></span></dt></dl></dd></dl></div><div class="section" title="Purpose"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="d0e533"></a>Purpose</h2></div></div></div><p>
    200. 

  200.         Evaluation of the performance of a DHCP server requires that it
    201. 

  201.         be tested under varying traffic loads.  perfdhcp is a testing
    202. 

  202.         tool with the capability to create traffic loads
    203. 

  203.         and generate statistics from the results.  Additional features,
    204. 

  204.         such as the ability to send customised DHCP packets, allow it to
    205. 

  205.         be used in a wide range of functional testing.
    206. 

  206.       </p></div><div class="section" title="Key features"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="perfdhcp-key-features"></a>Key features</h2></div></div></div><p>
    207. 

  207.         perfdhcp has a number of command line switches to
    208. 

  208.         control DHCP message exchanges. Currently they fall into
    209. 

  209.         the following categories:
    210. 

  210.         </p><div class="itemizedlist"><ul class="itemizedlist" type="disc"><li class="listitem"><p>
    211. 

  211.               Rate control - control how many DHCP exchanges
    212. 

  212.               are initiated within a period of time. The tool can also simulate
    213. 

  213.               best effort conditions by attempting to initiate as many DHCP
    214. 

  214.               packet exchanges as possible within a unit of time.
    215. 

  215.             </p></li><li class="listitem"><p>
    216. 

  216.               Test exit specifiers - control the conditions for test
    217. 

  217.               completion, including the number of initiated exchanges,
    218. 

  218.               the test period orthe maximum number of dropped packets.
    219. 

  219.             </p></li><li class="listitem"><p>
    220. 

  220.               Packet templates - specify files containing packet templates that
    221. 

  221.               are used by perfdhcp to create custom DHCP messages. The tool
    222. 

  222.               allows the specification of a number of values indicating
    223. 

  223.               offsets of values within a packet that are set by the tool.
    224. 

  224.             </p></li><li class="listitem"><p>
    225. 

  225.               Reporting - for each test produce a set of performance data
    226. 

  226.               including the achieved packet exchange rate (server performance).
    227. 

  227.               There are a number of diagnostic selectors available that
    228. 

  228.               enable periodic (intermediate) reporting, printing of packet timestamps,
    229. 

  229.               and the listing of detailed information about internal perfdhcp
    230. 

  230.               states (for debugging).
    231. 

  231.             </p></li><li class="listitem"><p>
    232. 

  232.               Different mode of operations - specify the DHCP protocol used
    233. 

  233.               (v4 or v6), two-way or four-way exchanges, use of the
    234. 

  234.               Rapid Commit option for DHCPv6.
    235. 

  235.             </p></li><li class="listitem"><p>
    236. 

  236.               IP layer options - specify the local/remote address, local interface
    237. 

  237.               and local port to be used for communication with DHCP server.
    238. 

  238.             </p></li></ul></div><p>
    239. 

  239.       </p></div><div class="section" title="Command line options"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="perfdhcp-command-line"></a>Command line options</h2></div></div></div><p>
    240. 

  240.         The following "help" output from the tool describes the
    241. 

  241.         command line switches.  This summary also lists the tool's
    242. 

  242.         possible exit codes as well as describing the
    243. 

  243.         error counters printed when the test is complete:
    244. 

  244.         </p><pre class="screen">$ <strong class="userinput"><code>./perfdhcp -h</code></strong>
    245. 

  245. perfdhcp [-hv] [-4|-6] [-r&lt;rate&gt;] [-t&lt;report&gt;] [-R&lt;range&gt;] [-b&lt;base&gt;]
    246. 

  246.     [-n&lt;num-request&gt;] [-p&lt;test-period&gt;] [-d&lt;drop-time&gt;] [-D&lt;max-drop&gt;]
    247. 

  247.     [-l&lt;local-addr|interface&gt;] [-P&lt;preload&gt;] [-a&lt;aggressivity&gt;]
    248. 

  248.     [-L&lt;local-port&gt;] [-s&lt;seed&gt;] [-i] [-B] [-c] [-1]
    249. 

  249.     [-T&lt;template-file&gt;] [-X&lt;xid-offset&gt;] [-O&lt;random-offset]
    250. 

  250.     [-E&lt;time-offset&gt;] [-S&lt;srvid-offset&gt;] [-I&lt;ip-offset&gt;]
    251. 

  251.     [-x&lt;diagnostic-selector&gt;] [-w&lt;wrapped&gt;] [server]
    252. 

  252. 

  253. The [server] argument is the name/address of the DHCP server to
    254. 

  254. contact.  For DHCPv4 operation, exchanges are initiated by
    255. 

  255. transmitting a DHCP DISCOVER to this address.
    256. 

  256. 

  257. For DHCPv6 operation, exchanges are initiated by transmitting a DHCP
    258. 

  258. SOLICIT to this address.  In the DHCPv6 case, the special name 'all'
    259. 

  259. can be used to refer to All_DHCP_Relay_Agents_and_Servers (the
    260. 

  260. multicast address FF02::1:2), or the special name 'servers' to refer
    261. 

  261. to All_DHCP_Servers (the multicast address FF05::1:3).  The [server]
    262. 

  262. argument is optional only in the case that -l is used to specify an
    263. 

  263. interface, in which case [server] defaults to 'all'.
    264. 

  264. 

  265. The default is to perform a single 4-way exchange, effectively pinging
    266. 

  266. the server.
    267. 

  267. The -r option is used to set up a performance test, without
    268. 

  268. it exchanges are initiated as fast as possible.
    269. 

  269. 

  270. Options:
    271. 

  271. -1: Take the server-ID option from the first received message.
    272. 

  272. -4: DHCPv4 operation (default). This is incompatible with the -6 option.
    273. 

  273. -6: DHCPv6 operation. This is incompatible with the -4 option.
    274. 

  274. -a&lt;aggressivity&gt;: When the target sending rate is not yet reached,
    275. 

  275.     control how many exchanges are initiated before the next pause.
    276. 

  276. -b&lt;base&gt;: The base mac, duid, IP, etc, used to simulate different
    277. 

  277.     clients.  This can be specified multiple times, each instance is
    278. 

  278.     in the &lt;type&gt;=&lt;value&gt; form, for instance:
    279. 

  279.     (and default) mac=00:0c:01:02:03:04.
    280. 

  280. -d&lt;drop-time&gt;: Specify the time after which a request is treated as
    281. 

  281.     having been lost.  The value is given in seconds and may contain a
    282. 

  282.     fractional component.  The default is 1 second.
    283. 

  283. -E&lt;time-offset&gt;: Offset of the (DHCPv4) secs field / (DHCPv6)
    284. 

  284.     elapsed-time option in the (second/request) template.
    285. 

  285.     The value 0 disables it.
    286. 

  286. -h: Print this help.
    287. 

  287. -i: Do only the initial part of an exchange: DO or SA, depending on
    288. 

  288.     whether -6 is given.
    289. 

  289. -I&lt;ip-offset&gt;: Offset of the (DHCPv4) IP address in the requested-IP
    290. 

  290.     option / (DHCPv6) IA_NA option in the (second/request) template.
    291. 

  291. -l&lt;local-addr|interface&gt;: For DHCPv4 operation, specify the local
    292. 

  292.     hostname/address to use when communicating with the server.  By
    293. 

  293.     default, the interface address through which traffic would
    294. 

  294.     normally be routed to the server is used.
    295. 

  295.     For DHCPv6 operation, specify the name of the network interface
    296. 

  296.     via which exchanges are initiated.
    297. 

  297. -L&lt;local-port&gt;: Specify the local port to use
    298. 

  298.     (the value 0 means to use the default).
    299. 

  299. -O&lt;random-offset&gt;: Offset of the last octet to randomize in the template.
    300. 

  300. -P&lt;preload&gt;: Initiate first &lt;preload&gt; exchanges back to back at startup.
    301. 

  301. -r&lt;rate&gt;: Initiate &lt;rate&gt; DORA/SARR (or if -i is given, DO/SA)
    302. 

  302.     exchanges per second.  A periodic report is generated showing the
    303. 

  303.     number of exchanges which were not completed, as well as the
    304. 

  304.     average response latency.  The program continues until
    305. 

  305.     interrupted, at which point a final report is generated.
    306. 

  306. -R&lt;range&gt;: Specify how many different clients are used. With 1
    307. 

  307.     (the default), all requests seem to come from the same client.
    308. 

  308. -s&lt;seed&gt;: Specify the seed for randomization, making it repeatable.
    309. 

  309. -S&lt;srvid-offset&gt;: Offset of the server-ID option in the
    310. 

  310.     (second/request) template.
    311. 

  311. -T&lt;template-file&gt;: The name of a file containing the template to use
    312. 

  312.     as a stream of hexadecimal digits.
    313. 

  313. -v: Report the version number of this program.
    314. 

  314. -w&lt;wrapped&gt;: Command to call with start/stop at the beginning/end of
    315. 

  315.     the program.
    316. 

  316. -x&lt;diagnostic-selector&gt;: Include extended diagnostics in the output.
    317. 

  317.     &lt;diagnostic-selector&gt; is a string of single-keywords specifying
    318. 

  318.     the operations for which verbose output is desired.  The selector
    319. 

  319.     keyletters are:
    320. 

  320.    * 'a': print the decoded command line arguments
    321. 

  321.    * 'e': print the exit reason
    322. 

  322.    * 'i': print rate processing details
    323. 

  323.    * 'r': print randomization details
    324. 

  324.    * 's': print first server-id
    325. 

  325.    * 't': when finished, print timers of all successful exchanges
    326. 

  326.    * 'T': when finished, print templates
    327. 

  327. -X&lt;xid-offset&gt;: Transaction ID (aka. xid) offset in the template.
    328. 

  328. 

  329. DHCPv4 only options:
    330. 

  330. -B: Force broadcast handling.
    331. 

  331. 

  332. DHCPv6 only options:
    333. 

  333. -c: Add a rapid commit option (exchanges will be SA).
    334. 

  334. 

  335. The remaining options are used only in conjunction with -r:
    336. 

  336. 

  337. -D&lt;max-drop&gt;: Abort the test if more than &lt;max-drop&gt; requests have
    338. 

  338.     been dropped.  Use -D0 to abort if even a single request has been
    339. 

  339.     dropped.  If &lt;max-drop&gt; includes the suffix '%', it specifies a
    340. 

  340.     maximum percentage of requests that may be dropped before abort.
    341. 

  341.     In this case, testing of the threshold begins after 10 requests
    342. 

  342.     have been expected to be received.
    343. 

  343. -n&lt;num-request&gt;: Initiate &lt;num-request&gt; transactions.  No report is
    344. 

  344.     generated until all transactions have been initiated/waited-for,
    345. 

  345.     after which a report is generated and the program terminates.
    346. 

  346. -p&lt;test-period&gt;: Send requests for the given test period, which is
    347. 

  347.     specified in the same manner as -d.  This can be used as an
    348. 

  348.     alternative to -n, or both options can be given, in which case the
    349. 

  349.     testing is completed when either limit is reached.
    350. 

  350. -t&lt;report&gt;: Delay in seconds between two periodic reports.
    351. 

  351. 

  352. Errors:
    353. 

  353. - tooshort: received a too short message
    354. 

  354. - orphans: received a message which doesn't match an exchange
    355. 

  355.    (duplicate, late or not related)
    356. 

  356. - locallimit: reached to local system limits when sending a message.
    357. 

  357. 

  358. Exit status:
    359. 

  359. The exit status is:
    360. 

  360. 0 on complete success.
    361. 

  361. 1 for a general error.
    362. 

  362. 2 if an error is found in the command line arguments.
    363. 

  363. 3 if there are no general failures in operation, but one or more
    364. 

  364.   exchanges are not successfully completed.
    365. 

  365. 

  366.         </pre><p>
    367. 

  367.       </p></div><div class="section" title="Starting perfdhcp"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="starting-perfdhcp"></a>Starting perfdhcp</h2></div></div></div><p>
    368. 

  368.         In order to run a performance test, at least two separate systems
    369. 

  369.         have to be installed: client and server. The first one must have
    370. 

  370.         perfdhcp installed, and the latter must be running the DHCP server
    371. 

  371.         (either v4 or v6). If only single system is available the client
    372. 

  372.         and server can be run on virtual machines (running on the same
    373. 

  373.         physical system) but in this case performance data may be heavily
    374. 

  374.         impacted by the overhead involved in running such the virtual
    375. 

  375.         machines.
    376. 

  376.       </p><p>
    377. 

  377.         Currently, perfdhcp is seen from the server perspective as relay agent.
    378. 

  378.         This simplifies its implementation: specifically there is no need to
    379. 

  379.         receive traffic sent to braodcast addresses. However, it does impose
    380. 

  380.         a requirement that the IPv4
    381. 

  381.         address has to be set manually on the interface that will be used to
    382. 

  382.         communicate with the server. For example, if the DHCPv4 server is listening
    383. 

  383.         on the interface connected to the 172.16.1.0 subnet, the interface on client
    384. 

  384.         machine has to have network address assigned from the same subnet, e.g.
    385. 

  385.         </p><pre class="screen"><strong class="userinput"><code>#ifconfig eth3 172.16.1.2. netmask 255.255.255.0 up</code></strong></pre><p>
    386. 

  386.       </p><p>
    387. 

  387.         As DHCP uses low port numbers (67 for DHCPv4 relays and
    388. 

  388.         547 for DHCPv6), running perfdhcp with non-root privileges will
    389. 

  389.         usually result in the error message similar to this:
    390. 

  390.         </p><pre class="screen"><strong class="userinput"><code>$./perfdhcp -4 -l eth3 -r 100 all</code></strong>
    391. 

  391. Error running perfdhcp: Failed to bind socket 3 to 172.16.1.2/port=67
    392. 

  392.         </pre><p>
    393. 

  393.         The '-L' command line switch allows the use of a custom local port.
    394. 

  394.         However, although the following command line will work:
    395. 

  395.         </p><pre class="screen"><strong class="userinput"><code>$./perfdhcp -4 -l eth3 -r 100 -L 10067 all</code></strong></pre><p>
    396. 

  396.         in the standard configuration no responses will be received
    397. 

  397.         from the DHCP server because the server responds to default relay
    398. 

  398.         port 67.  A way to overcome this issue is to run
    399. 

  399.         perfdhcp as root.
    400. 

  400.       </p></div><div class="section" title="perfdhcp command line examples"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="perfdhcp-commandline-examples"></a>perfdhcp command line examples</h2></div></div></div><p>
    401. 

  401.         In this section, a number of perfdhcp command line examples
    402. 

  402.         are presented as a quick start guide for new users. For the
    403. 

  403.         detailed list of command line options refer to
    404. 

  404.         <a class="xref" href="#perfdhcp-command-line" title="Command line options">the section called &#8220;Command line options&#8221;</a>.
    405. 

  405.       </p><div class="section" title="Example: basic usage"><div class="titlepage"><div><div><h3 class="title"><a name="perfdhcp-basic-usage"></a>Example: basic usage</h3></div></div></div><p>
    406. 

  406.           If server is listening on interface with IPv4 address 172.16.1.1,
    407. 

  407.           the simplest perfdhcp command line will look like:
    408. 

  408.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp 172.16.1.1</code></strong>
    409. 

  409. ***Rate statistics***
    410. 

  410. Rate: 206.345
    411. 

  411. 

  412. ***Statistics for: DISCOVER-OFFER***
    413. 

  413. sent packets: 21641
    414. 

  414. received packets: 350
    415. 

  415. drops: 21291
    416. 

  416. orphans: 0
    417. 

  417. 

  418. min delay: 9.022 ms
    419. 

  419. avg delay: 143.100 ms
    420. 

  420. max delay: 259.303 ms
    421. 

  421. std deviation: 56.074 ms
    422. 

  422. collected packets: 30
    423. 

  423. 

  424. ***Statistics for: REQUEST-ACK***
    425. 

  425. sent packets: 350
    426. 

  426. received packets: 268
    427. 

  427. drops: 82
    428. 

  428. orphans: 0
    429. 

  429. 

  430. min delay: 3.010 ms
    431. 

  431. avg delay: 152.470 ms
    432. 

  432. max delay: 258.634 ms
    433. 

  433. std deviation: 56.936 ms
    434. 

  434. collected packets: 0
    435. 

  435.           </pre><p>
    436. 

  436.           Here, perfdhcp uses remote address 172.16.1.1 as a
    437. 

  437.           destination address and will use a suitable local interface for
    438. 

  438.           communication. Since, no rate control parameters have been specified,
    439. 

  439.           it will initiate DHCP exchanges at the maximum possible rate. Due to the server's
    440. 

  440.           performance limitation, it is likely that many of the packets will be dropped.
    441. 

  441.           The performance test will continue running until it is
    442. 

  442.           interrupted by the user (with ^C).
    443. 

  443.         </p><p>
    444. 

  444.           The default performance statistics reported by perfdhcp have the
    445. 

  445.           following meaning:
    446. 

  446.           </p><div class="itemizedlist"><ul class="itemizedlist" type="disc"><li class="listitem"><p>Rate - number of packet exchanges (packet sent
    447. 

  447.             to the server and matching response received from the server)
    448. 

  448.             completed within a second.</p></li><li class="listitem"><p>sent packets - total number of DHCP packets of
    449. 

  449.             a specific type sent to the server.</p></li><li class="listitem"><p>received packets - total number of DHCP packets
    450. 

  450.             of specific type received from the server.</p></li><li class="listitem"><p>drops - number of dropped packets for the
    451. 

  451.             particular exchange. The number of dropped packets is calculated as
    452. 

  452.             the difference between the number of sent packets and number of
    453. 

  453.             response packets received from the server.  In some cases, the
    454. 

  454.             server will have sent a reponse but perfdhcp execution ended before
    455. 

  455.             the reponse arrived. In such case this packet will be counted
    456. 

  456.             as dropped.</p></li><li class="listitem"><p>orphans - number of packets that have been
    457. 

  457.             received from the server and did not match any packet sent by
    458. 

  458.             perfdhcp. This may occur if received packet has been sent
    459. 

  459.             to some other host or if then exchange timed out and
    460. 

  460.             the sent packet was removed from perfdhcp's list of packets
    461. 

  461.             awaiting a response.</p></li><li class="listitem"><p>min delay - minimum delay that occured between
    462. 

  462.             sending the packet to the server and receiving a reponse from
    463. 

  463.             it.</p></li><li class="listitem"><p>avg delay - average delay between sending the
    464. 

  464.             packet of the specific type the server and receiving a response
    465. 

  465.             from it.</p></li><li class="listitem"><p>max delay - maximum delay that occured between
    466. 

  466.             sending the packet to the server and receiving a response from
    467. 

  467.             it.</p></li><li class="listitem"><p>std deviation - standard deviation of the delay
    468. 

  468.             between sending the packet of a specific type to the server and
    469. 

  469.             receiving response from it.</p></li><li class="listitem"><p>collected packets - number of sent packets that
    470. 

  470.             were garbage collected. Packets may get garbage collected when
    471. 

  471.             the waiting time for server a response exceeds value set with the
    472. 

  472.             '-d' (drop time) switch.
    473. 

  473.             -d&lt;drop-time&gt;.</p></li></ul></div><p>
    474. 

  474.         </p><p>
    475. 

  475.           Note: should multiple interfaces on the system running perfdhcp be
    476. 

  476.           connected to the same subnet, the interface to be used for the test
    477. 

  477.           can be specified using either the interface name:
    478. 

  478.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp -l eth3</code></strong></pre><p>
    479. 

  479.           or a local address assigned to it:
    480. 

  480.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp -l 172.16.1.2</code></strong></pre><p>
    481. 

  481.         </p></div><div class="section" title="Example: rate control"><div class="titlepage"><div><div><h3 class="title"><a name="perfdhcp-rate-control"></a>Example: rate control</h3></div></div></div><p>
    482. 

  482.           In the examples above perfdhcp initiates new exchanges with a best
    483. 

  483.           effort rate. With this setting, many packets are expected to be dropped
    484. 

  484.           by the server due to performance limitations. In many cases though, it is
    485. 

  485.           desired to verify that the server can handle an expected (reasonable) rate
    486. 

  486.           without dropping any packets. The following command is an example of such
    487. 

  487.           a test: it causes perfdhcp to initiate 300 four-way exchanges
    488. 

  488.           per second, and runs the test for 60 seconds:
    489. 

  489.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp -l eth3 -p 60 -r 300</code></strong>
    490. 

  490. ***Rate statistics***
    491. 

  491. Rate: 256.683 exchanges/second, expected rate: 300 exchanges/second
    492. 

  492. 

  493. ***Statistics for: DISCOVER-OFFER***
    494. 

  494. sent packets: 17783
    495. 

  495. received packets: 15401
    496. 

  496. drops: 2382
    497. 

  497. orphans: 0
    498. 

  498. 

  499. min delay: 0.109 ms
    500. 

  500. avg delay: 75.756 ms
    501. 

  501. max delay: 575.614 ms
    502. 

  502. std deviation: 60.513 ms
    503. 

  503. collected packets: 11
    504. 

  504. 

  505. ***Statistics for: REQUEST-ACK***
    506. 

  506. sent packets: 15401
    507. 

  507. received packets: 15317
    508. 

  508. drops: 84
    509. 

  509. orphans: 0
    510. 

  510. 

  511. min delay: 0.536 ms
    512. 

  512. avg delay: 72.072 ms
    513. 

  513. max delay: 576.749 ms
    514. 

  514. std deviation: 58.189 ms
    515. 

  515. collected packets: 0
    516. 

  516.           </pre><p>
    517. 

  517.           Note that here, the packet drops for the DISCOVER-OFFER
    518. 

  518.           exchange have been significantly reduced (when compared with the
    519. 

  519.           output from the previous example) thanks to the setting of a
    520. 

  520.           reasonable rate. The non-zero number of packet drops and achieved
    521. 

  521.           rate (256/s) indicate that server's measured performance is lower than 300 leases
    522. 

  522.           per second. A further rate decrease should eliminate most of the packet
    523. 

  523.           drops and bring the achieved rate close to expected rate:
    524. 

  524.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp -l eth3 -p 60 -r 100 -R 30</code></strong>
    525. 

  525. ***Rate statistics***
    526. 

  526. Rate: 99.8164 exchanges/second, expected rate: 100 exchanges/second
    527. 

  527. 

  528. ***Statistics for: DISCOVER-OFFER***
    529. 

  529. sent packets: 5989
    530. 

  530. received packets: 5989
    531. 

  531. drops: 0
    532. 

  532. orphans: 0
    533. 

  533. 

  534. min delay: 0.023 ms
    535. 

  535. avg delay: 2.198 ms
    536. 

  536. max delay: 181.760 ms
    537. 

  537. std deviation: 9.429 ms
    538. 

  538. collected packets: 0
    539. 

  539. 

  540. ***Statistics for: REQUEST-ACK***
    541. 

  541. sent packets: 5989
    542. 

  542. received packets: 5989
    543. 

  543. drops: 0
    544. 

  544. orphans: 0
    545. 

  545. 

  546. min delay: 0.473 ms
    547. 

  547. avg delay: 2.355 ms
    548. 

  548. max delay: 189.658 ms
    549. 

  549. std deviation: 5.876 ms
    550. 

  550. collected packets: 0
    551. 

  551.           </pre><p>
    552. 

  552.           There are now no packet drops, confirming that the server is able to
    553. 

  553.           handle a load of 100 leases/second.
    554. 

  554.           Note that the last parameter (-R 30) configures perfdhcp to simulate
    555. 

  555.           traffic from 30 distinct clients.
    556. 

  556.         </p></div><div class="section" title="Example: templates"><div class="titlepage"><div><div><h3 class="title"><a name="perfdhcp-templates"></a>Example: templates</h3></div></div></div><p>
    557. 

  557.           By default the DHCP messages are formed with default options.  With
    558. 

  558.           template files, it is possible to define a custom packet format.
    559. 

  559.         </p><p>
    560. 

  560.           The template file names are specified with -T&lt;template-file&gt;
    561. 

  561.           command line option. This option can be specified zero, one or two times.
    562. 

  562.           The first occurence of this option refers to DISCOVER or SOLICIT message
    563. 

  563.           and the second occurence refers to REQUEST (DHCPv4 or DHCPv6) message.
    564. 

  564.           If -T option occurs only once the DISCOVER or SOLICIT message will be
    565. 

  565.           created from the template and the REQUEST message will be created
    566. 

  566.           dynamically (without the template). Note that each template file
    567. 

  567.           holds data for exactly one DHCP message type. Templates for multiple
    568. 

  568.           message types must not be combined in the single file.
    569. 

  569.           The content in template files is encoded in hexadecimal format.
    570. 

  570.           perfdhcp forms the packet by replacing parts of the binary stream read
    571. 

  571.           from the file with variable data such as elapsed time, hardware address, DUID
    572. 

  572.           etc. The offsets where such variable data is placed is specific to the
    573. 

  573.           template file and have to be specified from the command line. Refer to
    574. 

  574.           <a class="xref" href="#perfdhcp-command-line" title="Command line options">the section called &#8220;Command line options&#8221;</a> to find out how to
    575. 

  575.           specify offsets for particular options and fields. With the following
    576. 

  576.           command line the DHCPv6 SOLICIT and REQUEST packets will be formed from
    577. 

  577.           solicit.hex and request6.hex packets:
    578. 

  578.           </p><pre class="screen"><strong class="userinput"><code>#./perfdhcp -6 -l eth3 -r 100 -R 20 -T solicit.hex -T request6.hex -O 21 -E 84 -S 22 -I 40 servers</code></strong>
    579. 

  579. ***Rate statistics***
    580. 

  580. Rate: 99.5398 exchanges/second, expected rate: 100 exchanges/second
    581. 

  581. 

  582. ***Statistics for: SOLICIT-ADVERTISE***
    583. 

  583. sent packets: 570
    584. 

  584. received packets: 569
    585. 

  585. drops: 1
    586. 

  586. orphans: 0
    587. 

  587. 

  588. min delay: 0.259 ms
    589. 

  589. avg delay: 0.912 ms
    590. 

  590. max delay: 6.979 ms
    591. 

  591. std deviation: 0.709 ms
    592. 

  592. collected packets: 0
    593. 

  593. 

  594. ***Statistics for: REQUEST-REPLY***
    595. 

  595. sent packets: 569
    596. 

  596. received packets: 569
    597. 

  597. drops: 0
    598. 

  598. orphans: 0
    599. 

  599. 

  600. min delay: 0.084 ms
    601. 

  601. avg delay: 0.607 ms
    602. 

  602. max delay: 6.490 ms
    603. 

  603. std deviation: 0.518 ms
    604. 

  604. collected packets: 0
    605. 

  605.           </pre><p>
    606. 

  606.           where the switches have the following meaning:
    607. 

  607.           </p><div class="itemizedlist"><ul class="itemizedlist" type="disc"><li class="listitem"><p>two occurences of -O 21 - DUID's last octet
    608. 

  608.             positions in SOLICIT and REQUEST respectively.</p></li><li class="listitem"><p>-E 84 - elapsed time option position in the
    609. 

  609.             REQUEST template</p></li><li class="listitem"><p>-S 22 - server id position in the REQUEST
    610. 

  610.             template</p></li><li class="listitem"><p>-I 40 - IA_NA option position in the REQUEST
    611. 

  611.             template</p></li></ul></div><p>
    612. 

  612.         </p><p>
    613. 

  613.           The offsets of options indicate where they begin in the packet.
    614. 

  614.           The only exception from this rule is -O&lt;random-offset&gt;
    615. 

  615.           option that specifies the end of the DUID (DHCPv6) or MAC address
    616. 

  616.           (DHCPv4). Depending on the number of simulated clients
    617. 

  617.           (see -R&lt;random-range&gt; command line option) perfdhcp
    618. 

  618.           will be randomizing bytes in packet buffer starting from this
    619. 

  619.           position backwards. For the number of simulated clients
    620. 

  620.           &lt;= 256 only one octet (at random-offset position)
    621. 

  621.           will be ranomized, for the number of clients &lt;= 65536
    622. 

  622.           two octets (at random-offset and random-offset-1)
    623. 

  623.           will be randmized etc.
    624. 

  624.         </p></div></div></div></div></body></html>
    625.