hydra/docs/hydra.texi

\input texinfo   @c -*-texinfo-*-
@c %**start of header
@setfilename hydra.info
@settitle The Hydra HTTP Daemon
@set UPDATED Last Updated: 2 Jan 2001
@set COPYPHRASE Copyright @copyright{} 1996-2001 Jon Nelson and Larry Doolittle
@set VERSION $Revision: 1.2 $

@paragraphindent asis
@iftex
@parindent 0pt
@end iftex
@c @setchapternewpage odd
@c %**end of header

@iftex
@titlepage
@title The Hydra HTTP Daemon
@c @sp 2
@end iftex

@ifinfo
This file documents Hydra, an HTTP daemon for UN*X like machines.
@end ifinfo

@html
<h1 align="center">The Hydra HTTP Daemon</h1>
@end html

@ifinfo
@dircategory Networking
@direntry
* Hydra: (hydra).           The Hydra Webserver
@end direntry
@end ifinfo

@comment node-name,     next,           previous, up
@node Top, Introduction, , (dir)

Welcome to the documentation for Hydra, a high performance
HTTP Server for UN*X-alike computers, covered by the
@uref{Gnu_License,GNU General Public License}. This is a draft
document based on Boa documentation.
The on-line, updated copy of this documentation lives at
@uref{http://hydra.hellug.gr/,http://hydra.hellug.gr/}
@sp 1
@center @value{COPYPHRASE}
@center @value{UPDATED}, @value{VERSION}

@iftex
@end titlepage
@contents
@end iftex

@menu
* Introduction::
* Installation and Usage::
* Limits and Design Philosophy::
* Appendix::

 -- Detailed Node Listing --

Installation

* Files Used by Hydra::
* Compile-Time and Command-Line Options::
* hydra.conf Directives::
* Security::

Limits and Design Philosophy 

* Limits::
* Differences between Hydra and other web servers::
* Unexpected Behavior::

Appendix 

* License::
* Acknowledgments::
* Reference Documents::
* Other HTTP Servers::
* Benchmarks::
* Tools::
* Authors::

@end menu

@comment node-name,     next,           previous, up
@node    Introduction, Installation and Usage,top,top
@chapter Introduction

Hydra is a single-tasking HTTP server.  That means that unlike
traditional web servers, it does not fork for each incoming
connection, nor does it fork many copies of itself to handle multiple
connections.  It internally multiplexes all of the ongoing HTTP
connections, and forks only for CGI programs (which must be separate
processes), automatic directory generation, and automatic file
gunzipping.  Preliminary tests show Hydra is capable of
handling several thousand hits per second on a 300 MHz Pentium and
dozens of hits per second on a lowly 20 MHz 386/SX.

The primary design goals of Hydra are speed and security.  Security,
in the sense of @emph{can't be subverted by a malicious user,} not
@emph{fine grained access control and encrypted communications}.
Hydra is not intended as a feature-packed server; if you want one of those,
check out 
WN (@uref{http://hopf.math.nwu.edu/}) from John Franks.
Modifications to Hydra that improve its speed, security, robustness, and
portability, are eagerly sought.  Other features may be added if they
can be achieved without hurting the primary goals.

Hydra is based on Boa web server.
Boa was created in 1991 by Paul Phillips (@email{psp@@well.com}).
It is now being maintained and enhanced by Larry Doolittle
(@email{ldoolitt@@boa.org}) and Jon Nelson
(@email{jnelson@@boa.org}).
Please see the acknowledgement section for further
details. Hydra is maintained by
Nikos Mavroyanopoulos (@email{nmav@@gnutls.org}).


GNU/Linux is the development platform at the moment, other OS's are known to work.
If you'd like to contribute to this effort, contact Larry or Jon via e-mail.

@comment node-name,     next,           previous, up
@node    Installation and Usage, Limits and Design Philosophy, Introduction,top
@chapter Installation and Usage

Hydra is currently being developed and tested on GNU/Linux/i386.
The code is straightforward (more so than most other servers),
so it should run easily on most modern Unix-alike platforms.  Recent
versions of Hydra worked fine on FreeBSD, SunOS 4.1.4, GNU/Linux-SPARC,
and HP-UX 9.0.  Pre-1.2.0 GNU/Linux kernels may not work because of
deficient mmap() implementations.

@menu
* Installation::
* Files Used by Hydra::
* Compile-Time and Command-Line Options::
* Security::
@end menu

@comment node-name,     next,           previous, up
@node Installation,Files Used by Hydra,,Installation and Usage
@section Installation

@enumerate
 @item Unpack
 @enumerate
  @item Choose, and cd into, a convenient directory for the package.
  @item @kbd{tar -xvzf hydra-x.y.z.tar.gz}, or for those of you with an archaic
(non-GNU) tar; @kbd{gzip -cd &lt; hydra-x.y.z.tar.gz | tar -xvf -}
  @item Read the documentation.  Really.
 @end enumerate
 @item Build
  @enumerate
   @item cd into the @t{src} directory.
   @item (optional) Change the default SERVER_ROOT by setting the #define
    at the top of src/defines.h
   @item Type @kbd{./configure; make}
   @item Report any errors to the maintainers for resolution, or strike
    out on your own.
  @end enumerate
  @item Configure
  @enumerate 
   @item Choose a user and server port under which Hydra can run.  The
    traditional port is 80, and user @t{nobody} (create if
    you need to) is often a good selection for security purposes.
    If you don't have (or choose not to use) root privileges, you
    can not use port numbers less than 1024, nor can you switch user id.
   @item Choose a server root.  The @t{conf} directory within the
    server root must hold your copy of the configuration file
   @emph{hydra.conf}
   @item Choose locations for log files, CGI programs (if any), and
    the base of your URL tree.
   @item Set the location of the @t{mime.types} file.
   @item Edit @emph{conf/hydra.conf} according to your
    choices above (this file documents itself).  Read through this file
    to see what other features you can configure.
 @end enumerate
 @item Start
  @itemize
   @item Start Hydra.  If you didn't build the right SERVER_ROOT into the
    binary, you can specify it on the command line with the -c option
    (command line takes precedence).
    @example
    Example: ./hydra -c /usr/local/hydra
    @end example
  @end itemize
  
 @item Test
  @itemize
   @item At this point the server should run and serve documents.
    If not, check the error_log file for clues.
  @end itemize
  
 @item Install
  @itemize
   @item Copy the binary to a safe place, and put the invocation into
    your system startup scripts.  Use the same -c option you used
    in your initial tests.
  @end itemize
@end enumerate

@comment node-name,     next,           previous, up
@node    Files Used by Hydra, Compile-Time and Command-Line Options, Installation,Installation and Usage
@section Files Used by Hydra

@ftable @file
@item hydra.conf
  This file is the sole configuration file for Hydra. The directives in this 
  file are defined in the DIRECTIVES section. 
@item mime.types
  The MimeTypes <filename> defines what Content-Type Hydra will 
  send in an HTTP/1.0 or better transaction. 
  Set to /dev/null if you do not want to load a mime types file.
  Do *not* comment out (better use AddType!)
@end ftable

@comment node-name,     next,           previous, up
@node    Compile-Time and Command-Line Options, hydra.conf Directives, Files Used by Hydra,Installation and Usage
@section Compile-Time and Command-Line Options

@table @var
@item SERVER_ROOT
@itemx -c
 The default server root as #defined by @var{SERVER_ROOT} in 
 @file{defines.h} can be overridden on the commandline using the 
 @option{-c} option.  The server root must hold your local copy of the 
 configuration file @file{hydra.conf}.
 @example
  Example: /usr/sbin/hydra -c /etc/hydra
 @end example

@end table

@comment node-name,     next,           previous, up
@node    hydra.conf Directives, Security, Compile-Time and Command-Line Options, (top)
@section hydra.conf Directives

The Hydra configuration file is parsed with a lex/yacc or flex/bison 
generated parser. If it reports an error, the line number will be 
provided; it should be easy to spot. The syntax of each of these rules 
is very simple, and they can occur in any order. Where possible, these 
directives mimic those of NCSA httpd 1.3; I (Paul Phillips) saw no reason 
to introduce gratuitous differences. 

Note: the "ServerRoot" is not in this configuration file. It can be 
compiled into the server (see @file{defines.h}) or specified on the command
line with the @command{-c} option. 

The following directives are contained in the @file{hydra.conf} file, and most, 
but not all, are required. 

@table @option
 @item Port <Integer>
  This is the port that Hydra runs on. The default port for http servers is 80.
  If it is less than 1024, the server must be started as root.
 
 @item Listen <IP>
  The Internet address to bind(2) to, in quadded-octet form (numbers).  
  If you leave it out, it binds to all addresses (INADDR_ANY).
  
  The name you provide gets run through inet_aton(3), so you have to
  use dotted  quad notation.  This configuration is too important to trust some DNS.   
  
  You only get one "Listen" directive,  if you want service on multiple
  IP addresses, you have three choices:
  
  @enumerate
   @item Run hydra without a "Listen" directive:
    @itemize @bullet
    @item All addresses are treated the same; makes sense if the addresses 
     are localhost, ppp, and eth0.
    @item Use the VirtualHost directive below to point requests to different files.
     Should be good for a very large number of addresses (web hosting clients).
   @end itemize
    @item Run one copy of hydra per IP address:
    @itemize @bullet
     @item Each instance has its own configuration with its own
     "Listen" directive.  No big deal up to a few tens of addresses. Nice separation
     between clients.
    @end itemize
  @end enumerate
 
 @item User <username or UID>
 The name or UID the server should run as. For Hydra to attempt this, the
 server must be started as root. 
 
 @item Group <groupname or GID>
 The group name or GID the server should run as. For Hydra to attempt this,
 the server must be started as root. 
 
 @item ServerAdmin <email address>
 The email address where server problems should be sent. Note: this is not
 currently used.
 
 @item ErrorLog <filename>
 The location of the error log file. If this does not start with /, it is
 considered relative to the server root. Set to /dev/null if you don't want
 errors logged. 
 
 @item AccessLog <filename>
 The location of the access log file. If this does not start with /, it is
 considered relative to the server root. Comment out or set to /dev/null
 (less effective) to disable access logging. 
 
 @item VerboseCGILogs
 This is a logical switch and does not take any parameters. Comment out to
 disable. All it does is switch on or off logging of when CGIs are launched and when
 the children return.
 
 @item CgiLog <filename>
 The location of the CGI error log file.  If
 specified, this is the file that the stderr of CGIs is tied to.  Otherwise, writes
 to stderr meet the bit bucket.
 
 @item ServerName <server_name>
 The name of this server that should be sent back to clients if different
 than that returned by gethostname.
 
 @item VirtualHost
 This is a logical switch and does not take any parameters.
 Comment out to disable. Given  DocumentRoot /var/www, requests on interface `A' or
 IP `IP-A' become /var/www/IP-A. Example: http://localhost/ becomes 
 /var/www/127.0.0.1 
 
 @item DocumentRoot <directory>
 The root directory of the HTML documents. If this does not start with /, 
 it is considered relative to the server root. 
 
 @item UserDir <directory>
 The name of the directory which is appended onto a user's home directory 
 if a ~user request is received. 
 
 @item DirectoryIndex <filename>
 Name of the file to use as a pre-written HTML directory index. Please 
 make and use these files. On the fly creation of directory indexes
 can be slow. 
 
 @item DirectoryMaker <full pathname to program>
 Name of the program used
 to generate on-the-fly directory listings.  The program must take one or two
 command-line arguments, the first being the directory to index (absolute), and the
 second, which is  optional, should be the "title" of the document be. Comment out if
 you don't  want on the fly directory listings. If this  does not start with /, it is
 considered relative to the server root. 
 
 @item DirectoryCache <directory>
 DirectoryCache: If DirectoryIndex doesn't exist, and DirectoryMaker has been
 commented out, the the on-the-fly indexing of Hydra can be used  to generate indexes
 of directories. Be warned that the output is extremely minimal and can cause
 delays when slow disks are used.  Note: The DirectoryCache must be writable by the
 same user/group that Hydra runs as.
 
 @item KeepAliveMax <integer>
 Number of KeepAlive requests to allow per connection. Comment out, or set 
 to 0 to disable keepalive processing. 
 
 @item KeepAliveTimeout <integer>
 Number of seconds to wait before keepalive connections time out. 
 
 @item MimeTypes <file>
 The location of the mime.types file. If this does not start with /, it is 
 considered relative to the server root.
 Comment out to avoid loading mime.types (better use AddType!)
 
 @item DefaultType <mime type>
 MIME type used if the file extension is unknown, or there is no file 
 extension. 
 
 @item AddType <mime type> <extension> extension...
 Associates a MIME type 
 with an extension or extensions. 
 
 @item Redirect, Alias, and ScriptAlias
 Redirect, Alias, and ScriptAlias all have the same semantics -- 
 they match the beginning of a request and take appropriate action. 
 Use Redirect for other servers, Alias for the same server, and 
 ScriptAlias to enable directories for script execution.
 
 @item Redirect <path1> <path2>
  allows you to tell clients about documents which used to exist 
  in your server's namespace, but do not anymore. This allows you
  tell the clients where to look for the relocated document. 
  
 @item Alias <path1> <path2>
  aliases one path to another. Of course, symbolic links in the 
  file system work fine too. 
  
 @item ScriptAlias <path1> <path2>
  maps a virtual path to a directory for serving scripts. 

 @item Allow, Deny
 Only supported if Hydra is compiled with --enable-access-control.
 Allow and Deny allows pattern based access control using shell
 wildcards. The string the matching is performed on is the absolute
 filesystem filename. The Allow, Deny directives are processed in
 order until the first match is found, so to allow files containing
 the substring 123 but not the ones containing 1234 you would do:
     @itemize @bullet
         @item Deny *1234*
         @item Allow *123*
     @end itemize

 @item Allow <pattern>
 Allows files matching <pattern>

 @item Deny <pattern>
 Disallowes files matching <pattern>

@end table

@comment node-name,     next,           previous, up
@node    Security, , hydra.conf Directives, Installation and Usage
@section Security

Hydra has been designed to use the existing file system security.   In
@file{hydra.conf}, the directives @emph{user} and 
@emph{group} determine who Hydra will run as, if launched by root.
By default, the user/group is nobody/nogroup.  This allows quite a bit
of flexibility.  For example, if you want to disallow access to otherwise
accessible directories or files, simply make them inaccessible to 
nobody/nogroup. If the user that Hydra runs as is "hydra" and the groups that 
"hydra" belongs to include "web-stuff" then files/directories accessible 
by users with group "web-stuff" will also be accessible to Hydra.

The February 2000 hoo-rah from 
@uref{http://www.cert.org/advisories/CA-2000-02.html,CERT advisory CA-2000-02}
has little to do with Hydra.  As of version 0.94.4, Hydra's escaping rules have 
been cleaned up a little, but they weren't that bad before.  The example CGI
programs have been updated to show what effort is needed there.  If you
write, maintain, or use CGI programs under Hydra (or any other server) it's
worth your while to read and understand this advisory.  The real problem,
however, boils down to browser and web page designers emphasizing frills
over content and security.  The market leading browsers assume (incorrectly)
that all web pages are trustworthy.

@comment node-name,     next,           previous, up
@node    Limits and Design Philosophy,Appendix, Installation and Usage,top
@chapter Limits and Design Philosophy

There are many issues that become more difficult to resolve in a single 
tasking web server than in the normal forking model.  Here is a partial 
list -- there are probably others that haven't been encountered yet.

@menu
* Limits::
* Differences between Hydra and other web servers::
* Unexpected Behavior::
@end menu

@comment node-name,     next,           previous, up
@node    Limits,Differences between Hydra and other web servers,,Limits and Design Philosophy
@section Limits

@itemize @bullet
 @item Slow file systems
 
 The file systems being served should be much faster than the
 network connection to the HTTP requests, or performance will suffer.
 For instance, if a document is served from a CD-ROM, the whole server
 (including all other currently incomplete data transfers) will stall
 while the CD-ROM spins up.  This is a consequence of the fact that Hydra
 mmap()'s each file being served, and lets the kernel read and cache
 pages as best it knows how.  When the files come from a local disk
 (the faster the better), this is no problem, and in fact delivers
 nearly ideal performance under heavy load.  Avoid serving documents
 from NFS and CD-ROM unless you have even slower inbound net
 connections (e.g., POTS SLIP).
 
 @item DNS lookups
 
 Writing a nonblocking gethostbyaddr is a difficult and not very
 enjoyable task.  Paul Phillips experimented with several methods, 
 including a separate logging process, before removing hostname 
 lookups entirely. There is a companion program with Hydra 
 @file{util/resolver.pl} that will postprocess the logfiles and 
 replace IP addresses with hostnames, which is much faster no matter 
 what sort of server you run.
 
 @item Identd lookups
 
 Same difficulties as hostname lookups; not included.
 Hydra provides a REMOTE_PORT environment variable, in addition
 to REMOTE_ADDR, so that a CGI program can do its own ident.
 See the end of @t{examples/cgi-test.cgi}.

 @item Password file lookups via NIS
 
 If users are allowed to serve HTML from their home directories,
 password file lookups can potentially block the process.  To lessen
 the impact, each user's home directory is cached by Hydra so it need
 only be looked up once.
 
 @item Running out of file descriptors
 
 Since a file descriptor is needed for every ongoing connection
 (two for non-nph CGIs, directories, and automatic gunzipping of files),
 it is possible though highly improbable to run out of file
 descriptors.  The symptoms of this conditions may vary with
 your particular unix variant, but you will probably see log
 entries giving an error message for @t{accept}.
 Try to build your kernel to give an adequate number for
 your usage - GNU/Linux provides 256 out of the box, more than
 enough for most people.
@end itemize

@comment node-name,     next,           previous, up
@node    Differences between Hydra and other web servers,Unexpected Behavior,Limits,Limits and Design Philosophy
@section Differences between Hydra and other web servers

In the pursuit of speed and simplicity, some aspects of Hydra differ
from the popular web servers.  In no particular order:

@itemize @bullet
 @item @var{REMOTE_HOST} environment variable not set for CGI programs
 
  The @var{REMOTE_HOST} environment variable is not set for CGI programs,
  for reasons already described.  This is easily worked around because the
  IP address is provided in the @var{REMOTE_HOST} variable, so (if the CGI
  program actually cares) gethostbyaddr or a variant can be used.
  
 @item There are no server side includes (@acronym{SSI}) in Hydra
 
  We don't like them, and they are too slow to parse.  We will consider
  more efficient alternatives.
  
 @item There are no access control features
 
  Hydra will follow symbolic links, and serve any file that it can
  read.  The expectation is that you will configure Hydra to run as user
  "nobody", and only files configured world readable will come
  out.
  
 @item No chroot option
 
  There is no option to run chrooted.  If anybody wants this, and is
  willing to try out experimental code, contact the maintainers.
@end itemize

@comment node-name,     next,           previous, up
@node    Unexpected Behavior,,Differences between Hydra and other web servers,Limits and Design Philosophy
@section Unexpected Behavior

@itemize @bullet
 @item SIGHUP handling
 
 Like any good server, Hydra traps SIGHUP and rereads @file{hydra.conf}.
 However, under normal circumstances, it has already given away
 permissions, so many items listed in @file{hydra.conf} can not take effect.
 No attempt is made to change uid, gid, log files, or server port.
 All other configuration changes should take place smoothly.

 @item SIGUSR1 handling
 
 Hydra traps SIGUSR1 and prints useful statistics in the error log file.
 
 @item Relative URL handling
 
 Not all browsers handle relative URLs correctly.  Hydra will not
 cover up for this browser bug, and will typically report 404 Not Found
 for URL's containing odd combinations of "../" 's.
 
 Note: As of version 0.95.0 (unreleased) the URL parser has been 
 rewritten and *does* correctly handle relative URLs. 
@end itemize

@comment node-name,     next,           previous, up
@node     Appendix,,Limits and Design Philosophy,top
@appendix Appendix

@menu
* License::
* Acknowledgments::
* Reference Documents::
* Other HTTP Servers::
* Benchmarks::
* Tools::
* Authors::
@end menu

@comment node-name,     next,           previous, up
@node    License,Acknowledgments,,Appendix
@section License

This program is distributed under the
 @uref{http://www.gnu.org/copyleft/gpl.html,GNU General Public License}.
as noted in each source file:
@*

@smallexample
/*
 *  Hydra, an http server
 *  Copyright (C) 1995 Paul Phillips <psp@@well.com>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 1, or (at your option)
 *  any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 */
 
@end smallexample

@comment node-name,     next,           previous, up
@node    Acknowledgments,Reference Documents,License,Appendix
@section Acknowledgments

Paul Phillips wrote the first versions of Hydra, up to and including
version 0.91.  Version 0.92 of Hydra was officially released December 1996
by Larry Doolittle.  Version 0.93 was the development version of 0.94,
which was released in February 2000.

The Boa Webserver is currently (Feb 2000) maintained and enhanced by 
Larry Doolittle (@email{ldoolitt@@boa.org}) 
and Jon Nelson (@email{jnelson@@boa.org}).

We would like to thank Russ Nelson (@email{nelson@@crynwr.com})
for hosting the @uref{http://www.boa.org,web site}.

We would also like to thank Paul Philips for writing code that is
worth maintaining and supporting.

Many people have contributed to Hydra, including (but not
limited to) Charles F. Randall (@email{randall@@goldsys.com})
Christoph Lameter (@email{<chris@@waterf.org>}), 
Russ Nelson (@email{<nelson@@crynwr.com>}), Alain Magloire
(@email{<alain.magloire@@rcsm.ee.mcgill.ca>}),
and more recently, M. Drew Streib (@email{<dtype@@linux.com>}).

Paul Phillips records his acknowledgments as follows: 
@quotation
Thanks to everyone in the WWW community, in general a great bunch of people.
Special thanks to Clem Taylor (@email{<ctaylor@@eecis.udel.edu>}), who
provided invaluable feedback on many of my ideas, and offered good
ones of his own.  Also thanks to John Franks, author of wn, for
writing what I believe is the best webserver out there.
@end quotation

@comment node-name,     next,           previous, up
@node Reference Documents,Other HTTP Servers,Acknowledgments,Appendix
@section Reference Documents

Links to documents relevant to 
@uref{http://hydra.hellug.gr/,Hydra}
development and usage.  Incomplete, we're still working on this.
NCSA has a decent 
@uref{http://hoohoo.ncsa.uiuc.edu/docs/Library.html,page} along
these lines, too.

Also see Yahoo's List
@* @uref{http://www.yahoo.com/Computers_and_Internet/Software/Internet/World_Wide_Web/Servers/}

@itemize
 @item W3O HTTP page
 @* @uref{http://www.w3.org/pub/WWW/Protocols/}
 
 @item RFC 1945 HTTP-1.0 (informational)
 @* @uref{http://ds.internic.net/rfc/rfc1945.txt}

 @item IETF Working Group Draft 07 of HTTP-1.1 
 @* @uref{http://www.w3.org/pub/WWW/Protocols/HTTP/1.1/draft-ietf-http-v11-spec-07.txt}

 @item HTTP: A protocol for networked information
 @* @uref{http://www.w3.org/pub/WWW/Protocols/HTTP/HTTP2.html}

 @item The Common Gateway Interface (CGI)
 @* @uref{http://hoohoo.ncsa.uiuc.edu/cgi/overview.html}

 @item RFC 1738 URL syntax and semantics
 @* @uref{http://ds.internic.net/rfc/rfc1738.txt}

 @item RFC 1808 Relative URL syntax and semantics
 @* @uref{http://ds.internic.net/rfc/rfc1808.txt}
@end itemize

@comment node-name,     next,           previous, up
@node    Other HTTP Servers,Benchmarks,Reference Documents,Appendix
@section Other HTTP Servers

For unix-alike platforms, with published source code.

@itemize
 @item tiny/turbo/throttling httpd very similar to Hydra, with a throttling
 feature
 @* @uref{http://www.acme.com/software/thttpd/}

 @item Roxen: based on ulpc interpreter, non-forking (interpreter implements
 threading), GPL'd
 @* @uref{http://www.roxen.com/}

 @item WN: featureful, GPL'd
 @* @uref{http://hopf.math.nwu.edu/}

 @item Apache: fast, PD
 @* @uref{http://www.apache.org/}

 @item NCSA: standard, legal status?
 @* @uref{http://hoohoo.ncsa.uiuc.edu/}

 @item CERN: standard, PD, supports proxy
 @* @uref{http://www.w3.org/pub/WWW/Daemon/Status.html}

 @item xs-httpd 2.0: small, fast, pseudo-GPL'd
 @* @uref{http://www.stack.nl/~sven/xs-httpd/}

 @item bozohttpd.tar.gz sources, in perl
 @* @uref{ftp://ftp.eterna.com.au/bozo/bsf/attware/bozohttpd.tar.gz}

 @item Squid is actually an "Internet Object Cache"
 @* @uref{http://squid.nlanr.net/Squid/}
@end itemize

Also worth mentioning is Zeus.
It is commercial, with a free demo, so it doesn't belong on the list above.
Zeus seems to be based on technology similar to Hydra and thttpd,
but with more bells and whistles.
@* @uref{http://www.zeus.co.uk/products/server/} 

@comment node-name,     next,           previous, up
@node    Benchmarks,Tools,Other HTTP Servers,Appendix
@section Benchmarks

@itemize
@item ZeusBench (broken link)
@* @uref{http://www.zeus.co.uk/products/server/intro/bench2/zeusbench.shtml}

@item WebBench (binary-ware)
@* @uref{http://web1.zdnet.com/zdbop/webbench/webbench.html}

@item WebStone
@* @uref{http://www.mindcraft.com/benchmarks/webstone/}

@item SpecWeb96
@* @uref{http://www.specbench.org/osg/web96/}
@end itemize

@comment node-name,     next,           previous, up
@node    Tools,Authors,Benchmarks,Appendix
@section Tools

@itemize
@item Analog logfile analyzer
@* @uref{http://www.statslab.cam.ac.uk/@~sret1/analog/}

@item wwwstat logfile analyzer
@* @uref{http://www.ics.uci.edu/pub/websoft/wwwstat/}

@item gwstat wwwstat postprocessor
@* @uref{http://dis.cs.umass.edu/stats/gwstat.html} 

@item The Webalizer logfile analyzer
@* @uref{http://www.usagl.net/webalizer/}

@item cgiwrap
@* @uref{http://www.umr.edu/@~cgiwrap/}

@item suEXEC (Hydra would need to be ..umm.. "adjusted" to support this)
@* @uref{http://www.apache.org/docs/suexec.html}
@end itemize

Note: References last checked: 06 October 1997

@comment node-name,     next,           previous, up
@node    Authors,,Tools,Appendix
@section Authors

@itemize
 @item Conversion from linuxdoc SGML to texinfo by Jon Nelson
 @item Conversion to linuxdoc SGML by Jon Nelson
 @item Original HTML documentation by Larry Doolittle
 @item @value{COPYPHRASE}
@end itemize

@c variable
@c @printindex vr
@c concept
@c @printindex cp
@c function
@c @printindex fn
@c key
@c @printindex ky
@c program
@c @printindex pg
@c data type
@c @printindex tp

@bye