\chapter{Comparison of \MTA{}s}

% http://shearer.org/MTA_Comparison
% http://www.geocities.com/mailsoftware42/
% http://fanf.livejournal.com/50917.html
% http://archives.neohapsis.com/archives/postfix/2006-07/1762.html
% http://www.oreillynet.com/lpt/a/6849
% http://www.mailradar.com/mailstat/

\section{First release}
sendmail: 1983

postfix: 1999

qmail: 1996 (first beta 0.70), 1997 (first general 1.0)

exim: 1995

masqmail: 1999

exchange: 1993


\section{Lines of code (with sloccount on debian packages)}
sendmail: 93k

postfix: 92k

qmail: 18k

exim: 54k

masqmail: 14k

exchange: (no source available)


\section{Architecture}
sendmail: monolitic

postfix: modular

qmail: modular

exim: monolitic

masqmail: monolitic

exchange: (unknown)


\section{Design goals}
sendmail: flexibility

postfix: performance and security

qmail: security

exim: general, flexible \& extensive facilities for checking

masqmail: for non-permanent internet connection

exchange: groupware


\section{Market share (by Bernstein in 2001)}
sendmail: 42\%

postfix: 1.6\%

qmail: 17\%

exim: 1.6\%

masqmail: (unknown)

exchange: 18\%




1) complexity

2) security

3) simplicity of configuration and administration

4) flexibility of configuration and administration

5) code size

6) code quality

7) documentation (amount and quality)

8) community (amount and quality)

9) used it myself

10) had problems with it