<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">
<chapter>
<header>
<copyright>
<year>2002</year><year>2009</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
The contents of this file are subject to the Erlang Public License,
Version 1.1, (the "License"); you may not use this file except in
compliance with the License. You should have received a copy of the
Erlang Public License along with this software. If not, it can be
retrieved online at http://www.erlang.org/.
Software distributed under the License is distributed on an "AS IS"
basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
the License for the specific language governing rights and limitations
under the License.
</legalnotice>
<title>Performance comparison</title>
<prepared>Håkan Mattsson</prepared>
<responsible>Håkan Mattsson</responsible>
<docno></docno>
<approved>Håkan Mattsson</approved>
<checked></checked>
<date>2007-06-15</date>
<rev>%VSN%</rev>
<file>megaco_performance.xml</file>
</header>
<section>
<title>Comparison of encoder/decoders</title>
<p>The Megaco/H.248 standard defines both a plain text encoding and a
binary encoding (ASN.1 BER) and we have implemented encoders and
decoders for both. We do supply a bunch of different encoding/decoding
modules and the user may in fact implement their own (like our erl_dist
module). Using a non-standard encoding format has its obvious drawbacks,
but may be useful in some configurations.</p>
<p>We have made four different measurements of our Erlang/OTP
implementation of the Megaco/H.248 protocol stack, in order to compare
our different encoders/decoders. The result of each one is summarized
in the table below.</p>
<p>The result above are the fastest
of these configurations for each codec. The figures presented are
the average of all used messages.</p>
<p>For comparison, also included are performance figures
where the flex driver was built as <c>non-reentrant</c> flex
(figures within parenthesis). </p>
<table>
<row>
<cell align="left" valign="middle"><em>Codec and config</em></cell>
<cell align="center" valign="middle"><em>Size</em></cell>
<cell align="center" valign="middle"><em>Encode</em></cell>
<cell align="center" valign="middle"><em>Decode</em></cell>
<cell align="center" valign="middle"><em>Total</em></cell>
</row>
<!-- PRETTY -->
<row>
<cell align="left" valign="middle">pretty</cell>
<cell align="right" valign="middle">336</cell>
<cell align="right" valign="middle">22</cell>
<cell align="right" valign="middle">76</cell>
<cell align="right" valign="middle">98</cell>
</row>
<row>
<cell align="left" valign="middle">pretty [flex]</cell>
<cell align="right" valign="middle">336</cell>
<cell align="right" valign="middle">22 (22)</cell>
<cell align="right" valign="middle">41 (40)</cell>
<cell align="right" valign="middle">63 (62)</cell>
</row>
<!-- COMPACT -->
<row>
<cell align="left" valign="middle">compact</cell>
<cell align="right" valign="middle">181</cell>
<cell align="right" valign="middle">19</cell>
<cell align="right" valign="middle">63</cell>
<cell align="right" valign="middle">82</cell>
</row>
<row>
<cell align="left" valign="middle">compact [flex]</cell>
<cell align="right" valign="middle">181</cell>
<cell align="right" valign="middle">19 (19)</cell>
<cell align="right" valign="middle">38 (36)</cell>
<cell align="right" valign="middle">57 (55)</cell>
</row>
<!-- PER -->
<row>
<cell align="left" valign="middle">per bin</cell>
<cell align="right" valign="middle">91</cell>
<cell align="right" valign="middle">63</cell>
<cell align="right" valign="middle">69</cell>
<cell align="right" valign="middle">132</cell>
</row>
<row>
<cell align="left" valign="middle">per bin [driver]</cell>
<cell align="right" valign="middle">91</cell>
<cell align="right" valign="middle">43</cell>
<cell align="right" valign="middle">45</cell>
<cell align="right" valign="middle">88</cell>
</row>
<row>
<cell align="left" valign="middle">per bin [native]</cell>
<cell align="right" valign="middle">91</cell>
<cell align="right" valign="middle">47</cell>
<cell align="right" valign="middle">51</cell>
<cell align="right" valign="middle">99</cell>
</row>
<row>
<cell align="left" valign="middle">per bin [driver,native]</cell>
<cell align="right" valign="middle">91</cell>
<cell align="right" valign="middle">26</cell>
<cell align="right" valign="middle">29</cell>
<cell align="right" valign="middle">55</cell>
</row>
<!-- BER -->
<row>
<cell align="left" valign="middle">ber bin</cell>
<cell align="right" valign="middle">165</cell>
<cell align="right" valign="middle">35</cell>
<cell align="right" valign="middle">42</cell>
<cell align="right" valign="middle">77</cell>
</row>
<row>
<cell align="left" valign="middle">ber bin [driver]</cell>
<cell align="right" valign="middle">165</cell>
<cell align="right" valign="middle">35</cell>
<cell align="right" valign="middle">37</cell>
<cell align="right" valign="middle">72</cell>
</row>
<row>
<cell align="left" valign="middle">ber bin [native]</cell>
<cell align="right" valign="middle">165</cell>
<cell align="right" valign="middle">19</cell>
<cell align="right" valign="middle">26</cell>
<cell align="right" valign="middle">45</cell>
</row>
<row>
<cell align="left" valign="middle">ber bin [driver,native]</cell>
<cell align="right" valign="middle">165</cell>
<cell align="right" valign="middle">19</cell>
<cell align="right" valign="middle">20</cell>
<cell align="right" valign="middle">39</cell>
</row>
<!-- ERLANG -->
<row>
<cell align="left" valign="middle">erl_dist</cell>
<cell align="right" valign="middle">875</cell>
<cell align="right" valign="middle">5</cell>
<cell align="right" valign="middle">10</cell>
<cell align="right" valign="middle">15</cell>
</row>
<row>
<cell align="left" valign="middle">erl_dist [megaco_compressed]</cell>
<cell align="right" valign="middle">405</cell>
<cell align="right" valign="middle">6</cell>
<cell align="right" valign="middle">7</cell>
<cell align="right" valign="middle">13</cell>
</row>
<row>
<cell align="left" valign="middle">erl_dist [compressed]</cell>
<cell align="right" valign="middle">345</cell>
<cell align="right" valign="middle">86</cell>
<cell align="right" valign="middle">21</cell>
<cell align="right" valign="middle">107</cell>
</row>
<row>
<cell align="left" valign="middle">erl_dist [megaco_compressed,compressed]</cell>
<cell align="right" valign="middle">200</cell>
<cell align="right" valign="middle">71</cell>
<cell align="right" valign="middle">12</cell>
<cell align="right" valign="middle">83</cell>
</row>
<tcaption>Codec performance</tcaption>
</table>
</section>
<section>
<title>System performance characteristics</title>
<p>This is primarily a way to show the effects of using the
reentrant flex scanner instead of the non-reentrant. </p>
<p>As can be seen from the figures above there is no real difference
between a non-reentrant and an reentrant flex scanner when it
comes to the decode times of an individual message. </p>
<p>The real difference is instead in system characteristics, which
is best shown with the mstone1 test. </p>
<p>When running SMP erlang on a multi-core machine the "throughput"
is significantly higher. The mstone1 test is an extreme test,
but it shows what is gained by using the reentrant flex-scanner. </p>
<image file="mstone1.gif">
<icaption>MStone1 with mstone1.sh -d flex -s 8</icaption>
</image>
</section>
<section>
<title>Description of encoders/decoders</title>
<p>In Appendix A of the Megaco/H.248 specification (RFC 3525), there are
about 30 messages that shows a representative call flow. We have also
added a few extra version 1, version 2 and version 3 messages.
We have used these messages as basis for our measurements.
Our figures have not been weighted in regard to how frequent
the different kinds of messages that are sent between the media
gateway and its controller.</p>
<p>The test compares the following encoder/decoders:</p>
<list type="bulleted">
<item>
<p><em>pretty</em> - pretty printed text. In the text encoding,
the protocol stack implementors have the choice of using a
mix of short and long keywords. It is also possible to add
white spaces to improve readability. The pretty text encoding
utilizes long keywords and an indentation style like the
text examples in the Megaco/H.248 specification.</p>
</item>
<item>
<p><em>compact</em> - the compact text encoding uses the shortest
possible keywords and no optional white spaces.</p>
</item>
<item>
<p><em>ber</em> - ASN.1 BER.</p>
</item>
<item>
<p><em>per</em> - ASN.1 PER. Not standardized as a valid
Megaco/H.248 encoding, but included for the matter of completeness
as its encoding is extremely compact.</p>
</item>
<item>
<p><em>erl_dist</em> - Erlang's native distribution format. Not
standardized as a valid Megaco/H.248 encoding, but included
as a reference due to its well known performance characteristics.
Erlang is a dynamically typed language and any Erlang data
structure may be serialized to the erl_dist format by using
built-in functions.</p>
</item>
</list>
<p>The actual encoded messages have been collected in one directory per
encoding type, containing one file per encoded message.</p>
<p>Here follows an example of a text message to give a feeling of the
difference between the pretty and compact versions of text messages.
First the pretty printed, well indented version with long keywords:</p>
<p></p>
<pre>
MEGACO/1 [124.124.124.222]
Transaction = 9998 {
Context = - {
ServiceChange = ROOT {
\011Services {
\011 Method = Restart,
\011 ServiceChangeAddress = 55555,
\011 Profile = ResGW/1,
\011 Reason = "901 MG Cold Boot"
\011}
}
}
} </pre>
<p>Then the compact text version without indentation and with short
keywords:</p>
<pre>
!/1 [124.124.124.222] T=9998{
C=-{SC=ROOT{SV{MT=RS,AD=55555,PF=ResGW/1,RE="901 MG Cold Boot"}}}} </pre>
</section>
<section>
<title>Setup</title>
<p>The measurements has been performed on a
Dell PowerEdge 1950iii with
2* Intel Xeon L5430 @ 2.66 GHz, with 8 GB memory and
running SLES 10 SP2 x86_64, kernel 2.6.16.60-0.34-smp.
Software versions was open source OTP R13B and megaco-3.11.</p>
</section>
<section>
<title>Summary</title>
<p>In our measurements we have seen that there are no significant
differences in message sizes between ASN.1 BER and the compact
text format. Some care should be taken when using the pretty text
style (which is used in all the examples included in the protocol
specification and preferred during debugging sessions) since the
messages can then be quite large. If the message size really is a
serious issue, our per encoder should be used, as the ASN.1 PER
format is much more compact than all the other alternatives. Its
major drawback is that it is has not been approved as a valid
Megaco/H.248 message encoding.</p>
<p>When it comes to pure encode/decode performance, it turns out that:</p>
<list type="bulleted">
<item>
<p>our fastest binary encoder (ber) is about equal
to our fastest text encoder (compact). </p>
</item>
<item>
<p>our fastest binary decoder (ber) is about 47% (44%) faster than our
fastest text decoder (compact). </p>
</item>
</list>
<p>If the pure encode/decode performance really is a serious issue, our
erl_dist encoder could be used, as the encoding/decoding of the
erlang distribution format is much faster than all the other
alternatives. Its major drawback is that it is has not been approved
as a valid Megaco/H.248 message encoding.</p>
<p>There is no performance advantage of building (and using) a
non-reentrant flex scanner over a reentrant flex scanner (if flex
supports building such a scanner). </p>
<note>
<p>Please, observe that these performance figures are related to our
implementation in Erlang/OTP. Measurements of other implementations
using other tools and techniques may of course result in other
figures. </p>
</note>
</section>
</chapter>
