path: root/stasherpresent.tex

\pdfminorversion=4 % This is needed for impressive to work with this file (only functions when using pdflatex it seems)
\documentclass{beamer}
% \usetheme{Frankfurt}
\usetheme{Gopher}
\usecolortheme{Gopher}
\usepackage{german}
\usepackage{qtree}
\usepackage{graphicx}
%\usepackage{covington}
\usepackage{ulem}


\begin{document}

\title{stasher - Conceptionally exploring logstash in Go}
\author{Silvan Jegen} 
\date{\today}

\begin{frame}
\titlepage
\end{frame}

\begin{frame}
\frametitle{Contents} \tableofcontents
\end{frame}


\section{Logstash}
\subsection{What is it?}

\begin{frame}\frametitle{Uses}
	Extensible Markup Language (XML) \pause

	\begin{block}{XML aspects}
		\begin{itemize}
			\item Well-formedness
			\item Validation
			\item Namespaces
			\item Entities
		\end{itemize}
	\end{block}
\end{frame}

\begin{frame}\frametitle{XML in theory}
	\begin{block}{Related specifications}
		\begin{itemize}
			\item XSLT
			\item XPath
			\item XQuery
			\item XML Encryption
			\item ...
		\end{itemize}
	\end{block}
\end{frame}

\begin{frame}\frametitle{XML-based formats}
	\begin{block}{Variants}
		\begin{itemize}
				\item RDF XML
				\item XMPP
				\item EPUB
				\item XHTML
				\item ...
				\item 200+ more
		\end{itemize}
	\end{block}
\end{frame}

\subsection{XML in practice}
\begin{frame}\frametitle{XML in practice}
	\Huge XML in practice
\end{frame}

\begin{frame}\frametitle{XML in practice}
	\begin{block}{Enterprise usage}
		\begin{itemize}
			\item SOAP
			\item Configuration
			\item Data storage/exchange
			\item Java ecosystem...
		\end{itemize}
	\end{block}
\end{frame}

\begin{frame}[fragile]\frametitle{Markup}

	Annotate parts of text with additional information \pause

	\begin{block}{Text Markup}
		\begin{verbatim}
				<document>Some text <tag>some other text that
				should be tagged</tag> even <tag2>more</tag2>
				text...</document>
		\end{verbatim}
	\end{block}
\end{frame}

\begin{frame}[fragile]\frametitle{XML vs. JSON}
	\begin{block}{XML}
		\begin{verbatim}
				<document>Some text <tag>some other text that
				should be tagged</tag> even <tag2>more</tag2>
				text...</document>
		\end{verbatim}
	\end{block}  \pause
	\begin{block}{JSON}
		\begin{verbatim}
			["Some text ", {"t": "tag", "s": "some other text
			that should be tagged"}, "even", {"t": "tag2",
			"s": "more"}, "text..."]
		\end{verbatim}
	\end{block}
\end{frame}

\section{Dealing with XML}
\begin{frame}\frametitle{Dealing with XML}
	\Huge Dealing with XML
\end{frame}

\subsection{Programming interfaces}
\begin{frame}\frametitle{Programming interfaces}
	\begin{itemize}
		\item Stream-oriented (SAX, Stax)
		\item Tree traversal (DOM)
		\item XML Data binding
		\item Transformation languages (XSLT, XQuery) \pause
		\item Other?
	\end{itemize}
\end{frame}

\subsection{Benchmark}
\begin{frame}\frametitle{Benchmark}

	\begin{itemize}
		\item ezxml
		\item Golang encoding/xml
		\item mxml  ('Mini-XML', not 'Minimal XML')
		\item Python 2 ElementTree
		\item sxmlc
		\item yxml
	\end{itemize}

	All code available at:

	git://git.sillymon.ch/slcon3.git

\end{frame}

\begin{frame}\frametitle{Benchmark setup}
	Linux machine, i7 CPU, 8GB RAM
	\begin{itemize}
		\item 627MB of XML in 10'000 files from PubMed Central
		\item Printing the article titles
		\item 20 runs (after cache warming)
		\item Single-threaded (except for Go)
	\end{itemize}
\end{frame}

\begin{frame}[fragile]\frametitle{ezxml}
	\begin{block}{ezxml}
		\begin{itemize}
			\item Program: 21 lines
			\item Library: 623 lines
			\item URL: http://ezxml.sourceforge.net/
			\item Type: DOM (Level 3; XPath)
		\end{itemize}
	\end{block}
	\begin{verbatim}
		ezxml_t title = ezxml_get(ezdoc, "front", 0,\
		  "article-meta", 0, "title-group", 0,\
		  "article-title", -1);
	\end{verbatim}
\end{frame}

\begin{frame}[fragile]\frametitle{Go encoding/xml}
	\begin{block}{Go encoding/xml}
		\begin{itemize}
			\item Program: 30 lines
			\item Library: 6235 lines (stdlib)
			\item URL: https://golang.org/pkg/encoding/xml/
			\item Type: XML Data binding
		\end{itemize}
	\end{block}
	\begin{verbatim}
		type article struct {
		  Title string `xml:"front>article-meta>title-group>\
		                article-title"`
		}
	\end{verbatim}
\end{frame}

\begin{frame}[fragile]\frametitle{mxml}
	\begin{block}{mxml}
		\begin{itemize}
			\item Program: 38 lines
			\item Library: 9633 lines
			\item URL: http://www.minixml.org/
			\item Type: DOM (Level 3; Xpath)?
		\end{itemize}
	\end{block}
	\begin{verbatim}
		node = mxmlFindElement(root, root, "title-group",
		 NULL, NULL, MXML_DESCEND);
	\end{verbatim}
\end{frame}

\begin{frame}[fragile]\frametitle{Python 2 ElementTree}
	\begin{block}{Python 2 ElementTree}
		\begin{itemize}
			\item Program: 12 lines
			\item Library: 1107 lines (stdlib)
			\item URL:

				    https://docs.python.org/2/library/xml.etree.element\-tree.html

			\item Type: DOM (Level 3; Xpath)?
		\end{itemize}
	\end{block}
	\begin{verbatim}
		tg = r.findall("./front/article-meta/title-group")
		at = tg[0].find("article-title")
	\end{verbatim}
\end{frame}

\begin{frame}[fragile]\frametitle{sxmlc}
	\begin{block}{sxmlc}
		\begin{itemize}
			\item Program: 59 lines
			\item Library: 2690 lines
			\item URL: http://sxmlc.sourceforge.net/
			\item Type: DOM
		\end{itemize}
	\end{block}
	\begin{verbatim}
		const char *path[] = {"front", "article-meta",\
		           "title-group", "article-title", NULL};
		for (int i = 0; path[i]; i++) {
		  next = find_child_node(next, path[i]);
		  if (!next) {
		    fprintf(stderr, "Could not find '%s'
				                 tag.\n", path[i]);
		    return;
		  }
		}
	\end{verbatim}
\end{frame}

\begin{frame}[fragile]\frametitle{yxml}
	\begin{block}{yxml}
		\begin{itemize}
			\item Program: 103 lines
			\item Library: 1039 lines
			\item URL: https://dev.yorhel.nl/yxml
			\item Type: Stream-oriented
		\end{itemize}
	\end{block}
	\begin{verbatim}
		case YXML_ELEMSTART:
		 if (!strcmp(state->elem, "title-group")) {
		    intitlegroup = 1;
		 } else if (!strcmp(state->elem, "article-title")
		            && intitlegroup) {
		   printf("%s: ", state->elem);
		   inarticletitle = 1;
		 }
		 break;
	\end{verbatim}
\end{frame}


\begin{frame}\frametitle{Time \& Size}

	\begin{tabular}{ l | c | c | c | c | c }
						            &      mean  &      $\sigma$  &      min  &      max &     size   \\     \hline
		Python ElementTree  &     155.5  &         8.869  &    145.7  &    172.3 &    N/A   \\
		Go encoding/xml     &     48.34  &         4.982  &    35.33  &    52.92 &    2M   \\
		mxml                &     23.96  &         1.841  &    22.32  &    27.64 &    N/A   \\
		sxmlc               &     15.51  &         0.259  &    15.12  &    16.01 &    41K    \\
		ezxml               &     6.460  &         0.058  &    6.366  &    6.592 &    31K   \\
		yxml                &     4.123  &         0.220  &    3.885  &    4.520 &    18K
	\end{tabular}

\end{frame}


\section{Conclusion}
\begin{frame}[fragile]\frametitle{Conclusion}
	\begin{itemize}
		\item Complex specifications, (comparatively) hard to parse and verbose
		\item Use the ezxml or yxml libraries \pause
		\item Ok:
					\begin{verbatim}
						<document>Some text <tag>some other text that
						should be tagged</tag> even <tag2>more</tag2>
						text...</document>
					\end{verbatim} \pause
		\item No:
					\begin{verbatim}
						<thing><key>Key</key><value>Val</value></thing>
					\end{verbatim}
	\end{itemize}
\end{frame}


\begin{frame}\frametitle{Thanks for your attention}

	Questions?

\end{frame}

\end{document}