Common Gateway Interface
In computing, Common Gateway Interface is an interface specification for web servers to execute programs like console applications running on a server that generates web pages dynamically. Such programs are known as CGI scripts or simply as CGIs. The specifics of how the script is executed by the server are determined by the server. In the common case, a CGI script executes at the time a request is made and generates HTML.
In brief, an HTTP GET or POST request from the client may send HTML form data to the CGI program via standard input. Other data, such as URL paths, and HTTP header data, are presented as process environment variables.
History
In 1993 the National Center for Supercomputing Applications team wrote the specification for calling command line executables on the www-talk mailing list. The other Web server developers adopted it, and it has been a standard for Web servers ever since. A work group chaired by Ken Coar started in November 1997 to get the NCSA definition of CGI more formally defined. This work resulted in RFC 3875, which specified CGI Version 1.1. Specifically mentioned in the RFC are the following contributors:- Rob McCool
- John Franks
- Ari Luotonen
- Tony Sanders
- George Phillips
The name CGI comes from the early days of the web, where webmasters wanted to connect legacy information systems such as databases to their web servers. The CGI program was executed by the server that provided a common "gateway" between the web server and the legacy information system.
Purpose of the CGI specification
Each web server runs HTTP server software, which responds to requests from web browsers. Generally, the HTTP server has a directory, which is designated as a document collection — files that can be sent to Web browsers connected to this server. For example, if the Web server has the domain nameexample.com
, and its document collection is stored at /usr/local/apache/htdocs
in the local file system, then the Web server will respond to a request for http://example.com/index.html
by sending to the browser the file /usr/local/apache/htdocs/index.html
.For pages constructed on the fly, the server software may defer requests to separate programs and relay the results to the requesting client. In the early days of the web, such programs were usually small and written in a scripting language; hence, they were known as scripts.
Such programs usually require some additional information to be specified with the request. For instance, if Wikipedia were implemented as a script, one thing the script would need to know is whether the user is logged in and, if logged in, under which name. The content at the top of a Wikipedia page depends on this information.
HTTP provides ways for browsers to pass such information to the web server, e.g. as part of the URL. The server software must then pass this information through to the script somehow.
Conversely, upon returning, the script must provide all the information required by HTTP for a response to the request: the HTTP status of the request, the document content, the document type, et cetera.
Initially, different server software would use different ways to exchange this information with scripts. As a result, it wasn't possible to write scripts that would work unmodified for different server software, even though the information being exchanged was the same. Therefore, it was decided to specify a way for exchanging this information: CGI.
Webpage generating programs invoked by server software that operate according to the CGI specification are known as CGI scripts.
This specification was quickly adopted and is still supported by all well-known server software, such as Apache, IIS, and node.js-based servers.
An early use of CGI scripts was to process forms. In the beginning of HTML, HTML forms typically had an "action" attribute and a button designated as the "submit" button. When the submit button is pushed the URI specified in the "action" attribute would be sent to the server with the data from the form sent as a query string. If the "action" specifies a CGI script then the CGI script would be executed and it then produces an HTML page.
Using CGI scripts
A web server allows its owner to configure which URLs shall be handled by which CGI scripts.This is usually done by marking a new directory within the document collection as containing CGI scripts — its name is often
cgi-bin
. For example, /usr/local/apache/htdocs/cgi-bin
could be designated as a CGI directory on the web server. When a Web browser requests a URL that points to a file within the CGI directory, then, instead of simply sending that file to the Web browser, the HTTP server runs the specified script and passes the output of the script to the Web browser. That is, anything that the script sends to standard output is passed to the Web client instead of being shown on-screen in a terminal window.As remarked above, the CGI specification defines how additional information passed with the request is passed to the script.
For instance, if a slash and additional directory name are appended to the URL immediately after the name of the script, then that path is stored in the
PATH_INFO
environment variable before the script is called. If parameters are sent to the script via an HTTP GET request, then those parameters are stored in the QUERY_STRING
environment variable before the script is called. If parameters are sent to the script via an HTTP POST request, they are passed to the script's standard input. The script can then read these environment variables or data from standard input and adapt to the Web browser's request.Example
The following Perl program shows all the environment variables passed by the Web server:- !/usr/bin/env perl
printenv — a CGI program that just prints its environment
=cut
print "Content-Type: text/plain\n\n";
for my $var
If a Web browser issues a request for the environment variables at
http://example.com/cgi-bin/printenv.pl/foo/bar?var1=value1&var2=with%20percent%20encoding
, a 64-bit Windows 7 web server running cygwin returns the following information:COMSPEC="C:\Windows\system32\cmd.exe"
DOCUMENT_ROOT="C:/Program Files /Apache Software Foundation/Apache2.4/htdocs"
GATEWAY_INTERFACE="CGI/1.1"
HOME="/home/SYSTEM"
HTTP_ACCEPT="text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8"
HTTP_ACCEPT_CHARSET="ISO-8859-1,utf-8;q=0.7,*;q=0.7"
HTTP_ACCEPT_ENCODING="gzip, deflate, br"
HTTP_ACCEPT_LANGUAGE="en-us,en;q=0.5"
HTTP_CONNECTION="keep-alive"
HTTP_HOST="example.com"
HTTP_USER_AGENT="Mozilla/5.0 Gecko/20100101 Firefox/67.0"
PATH="/home/SYSTEM/bin:/bin:/cygdrive/c/progra~2/php:/cygdrive/c/windows/system32:..."
PATHEXT=".COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC"
PATH_INFO="/foo/bar"
PATH_TRANSLATED="C:\Program Files \Apache Software Foundation\Apache2.4\htdocs\foo\bar"
QUERY_STRING="var1=value1&var2=with%20percent%20encoding"
REMOTE_ADDR="127.0.0.1"
REMOTE_PORT="63555"
REQUEST_METHOD="GET"
REQUEST_URI="/cgi-bin/printenv.pl/foo/bar?var1=value1&var2=with%20percent%20encoding"
SCRIPT_FILENAME="C:/Program Files /Apache Software Foundation/Apache2.4/cgi-bin/printenv.pl"
SCRIPT_NAME="/cgi-bin/printenv.pl"
SERVER_ADDR="127.0.0.1"
SERVER_ADMIN=""
SERVER_NAME="127.0.0.1"
SERVER_PORT="80"
SERVER_PROTOCOL="HTTP/1.1"
SERVER_SIGNATURE=""
SERVER_SOFTWARE="Apache/2.4.39 PHP/7.3.7"
SYSTEMROOT="C:\Windows"
TERM="cygwin"
WINDIR="C:\Windows"
Some, but not all, of these variables are defined by the CGI standard.
Some, such as
PATH_INFO
, QUERY_STRING
, and the ones starting with HTTP_
, pass information along from the HTTP request.From the environment, it can be seen that the Web browser is Firefox running on a Windows 7 PC, the Web server is Apache running on a system that emulates Unix, and the CGI script is named
cgi-bin/printenv.pl
.The program could then generate any content, write that to standard output, and the Web server will transmit it to the browser.
The following are environment variables passed to CGI programs:
- Server specific variables:
- *
SERVER_SOFTWARE
: name/version of HTTP server. - *
SERVER_NAME
: host name of the server, may be dot-decimal IP address. - *
GATEWAY_INTERFACE
: CGI/version. - Request specific variables:
- *
SERVER_PROTOCOL
: HTTP/version. - *
SERVER_PORT
: TCP port. - *
REQUEST_METHOD
: name of HTTP method. - *
PATH_INFO
: path suffix, if appended to URL after program name and a slash. - *
PATH_TRANSLATED
: corresponding full path as supposed by server, ifPATH_INFO
is present. - *
SCRIPT_NAME
: relative path to the program, like/cgi-bin/script.cgi
. - *
QUERY_STRING
: the part of URL after ? character. The query string may be composed of *name=value pairs separated with ampersands when used to submit form data transferred via GET method as defined by HTML application/x-www-form-urlencoded. - *
REMOTE_HOST
: host name of the client, unset if server did not perform such lookup. - *
REMOTE_ADDR
: IP address of the client. - *
AUTH_TYPE
: identification type, if applicable. - *
REMOTE_USER
used for certainAUTH_TYPE
s. - *
REMOTE_IDENT
: see ident, only if server performed such lookup. - *
CONTENT_TYPE
: Internet media type of input data if PUT or POST method are used, as provided via HTTP header. - *
CONTENT_LENGTH
: similarly, size of input data if provided via HTTP header. - * Variables passed by user agent contain values of corresponding HTTP headers and therefore have the same sense.
The header is encoded in the same way as an HTTP header and must include the MIME type of the document returned. The headers, supplemented by the Web server, are generally forwarded with the response back to the user.
Here is a simple CGI program written in Python 3 along with the HTML that handles a simple addition problem.
add.html
:add.cgi
:- !/usr/bin/env python3
cgitb.enable
input_data = cgi.FieldStorage
print # HTML is following
print # Leave a blank line
try:
num1 = int
num2 = int
except:
raise SystemExit
Deployment
A Web server that supports CGI can be configured to interpret a URL that it serves as a reference to a CGI script. A common convention is to have acgi-bin/
directory at the base of the directory tree and treat all executable files within this directory as CGI scripts. Another popular convention is to use filename extensions; for instance, if CGI scripts are consistently given the extension .cgi
, the web server can be configured to interpret all such files as CGI scripts. While convenient, and required by many prepackaged scripts, it opens the server to attack if a remote user can upload executable code with the proper extension.In the case of HTTP PUT or POSTs, the user-submitted data are provided to the program via the standard input. The Web server creates a subset of the environment variables passed to it and adds details pertinent to the HTTP environment.
Uses
CGI is often used to process input information from the user and produce the appropriate output. An example of a CGI program is one implementing a wiki. The user agent requests the name of an entry; the Web server executes the CGI; the CGI program retrieves the source of that entry's page, transforms it into HTML, and prints the result. The web server receives the input from the CGI and transmits it to the user agent. If the "Edit this page" link is clicked, the CGI populates an HTMLtextarea
or other editing control with the page's contents, and saves it back to the server when the user submits the form in it.Security
CGI programs run, by default, in the security context of the web server. When first introduced a number of example scripts were provided with the reference distributions of the NCSA, Apache and CERN web servers to show how shell scripts or C programs could be coded to make use of the new CGI. One such example script was a CGI program called PHF that implemented a simple phone book.In common with a number of other scripts at the time, this script made use of a function: escape_shell_cmd. The function was supposed to sanitize its argument, which came from user input and then pass the input to the Unix shell, to be run in the security context of the web server. The script did not correctly sanitize all input and allowed new lines to be passed to the shell, which effectively allowed multiple commands to be run. The results of these commands were then displayed on the web server. If the security context of the web server allowed it, malicious commands could be executed by attackers.
This was the first widespread example of a new type of web based attack, where unsanitized data from web users could lead to execution of code on a web server. Because the example code was installed by default, attacks were widespread and led to a number of security advisories in early 1996.
Alternatives
Calling a command generally means the invocation of a newly created process on the server. Starting the process can consume much more time and memory than the actual work of generating the output, especially when the program still needs to be interpreted or compiled.If the command is called often, the resulting workload can quickly overwhelm the server.
The overhead involved in process creation can be reduced by techniques such as FastCGI that "prefork" interpreter processes, or by running the application code entirely within the web server, using extension modules such as mod_perl or mod_php. Another way to reduce the overhead is to use precompiled CGI programs, e.g. by writing them in languages such as C or C++, rather than interpreted or compiled-on-the-fly languages such as Perl or PHP, or by implementing the page generating software as a custom webserver module.
Alternative approaches include:
- Extensions such as Apache modules, NSAPI plugins, and ISAPI plugins allow third-party software to run on the web server. Web 2.0 allows to transfer data from the client to the server without using HTML forms and without the user noticing.
- FastCGI reduces overhead by allowing a single, long-running process to handle more than one user request. Unlike converting an application to a web server plug-in, FastCGI applications remain independent of the web server.
- Simple Common Gateway Interface or SCGI is designed to be easier to implement, yet it reduces latency in some operations compared to CGI.
- Replacement of the architecture for dynamic websites can also be used. This is the approach taken by Java EE, which runs Java code in a Java servlet container in order to serve dynamic content and optionally static content. This approach replaces the overhead of generating and destroying processes with the much lower overhead of generating and destroying threads, and also exposes the programmer to the library that comes with Java Platform, Standard Edition on which the version of Java EE in use is based.