WURFL InFuze Module for NGINX Plus: User Guide

This document is aimed at developers and system administrators who intend to install and configure the WURFL InFuze Module for NiGINX Plus on Unix, Linux, and other Unix-based systems.

Note: Unlike Open Source NGINX, you do not need to compile and link the NGINX Plus module.

Installing libwurfl

In order for the Module to work it is ESSENTIAL that the libwurfl library is installed on your system. libwurfl is provided in your Customer Vault/FileX.

If you have not already installed libwurfl, instructions can be found here. Release notes for each API can be found here.

Installing NGINX Plus

Please follow these instructions to install NGINX Plus for your platform.

Module Installation Guide

Login to your ScientiaMobile account to acess your Customer Vault, download your package, unpack it, and select the correct .so library for your Platform/NGINX Plus combination.

Tip: In the following example, W.X.Y.Z refers to WURFL module version and A.B.C. refers to NGINX Plus version.

Copy the .so module library to the /etc/nginx/modules folder.

$ sudo cp ngx_scientiamobile_http_wurfl-W.X.Y.Z_module_A.B.C.so  /etc/nginx/modules

Make a symbolic link for the .so module library in your /etc/nginx/modules folder.

$ sudo ln -s /etc/nginx/modules/ngx_scientiamobile_http_wurfl-W.X.Y.Z_module_A.B.C.so /etc/nginx/modules/ngx_http_wurfl_module.so

WURFL Data Snapshot

To perform lookups, you will need a copy of your WURFL data snapshot (also referred to as the wurfl.xml). While there is one included in the release package, it is intended to be a sample and will not contain all of your licensed capabilities. Your licensed WURFL data snapshot can be accessed by following these directions.

Module Configuration Guide

Below is an example nginx.conf configuration file for WURFL setup. You may find a sample .conf file in your NGINX Plus module package.
Please refer to the Module Command Table below which explains each element in detail (Table 1), their parameters, constraints, and default recommended settings.

# ...

# -- Uncomment this if you are using NGINX Plus
# -- or you compiled WURFL module with --add-dynamic-module (WURFL API version 1.8.1.0 or above / NGINX OSS 1.9.11 or above).
load_module modules/ngx_http_wurfl_module.so;

# ...

http {
    # ...

    # -- Command to enable WURFL module. Since WURFL module version 1.8.4.1, WURFL module is DISABLED BY DEFAULT.
    # -- Used to switch on/off the WURFL module; if set 'off', NGINX will ignore all wurfl_* configuration.
    # -- If the WURFL module is disabled, NGINX variables used to handle WURFL detection results (i.e: $wurfl_id , $wurfl_cap_is_smarttv ... ) are still valid but their values will be "empty string".
    # -- Valid values are on/off. Default value is off.
    wurfl_enable    on;

    # -- WURFL root definition, one per config. User MUST specify this path in order to make WURFL engine correctly start.
    wurfl_root      /usr/share/wurfl/wurfl.zip;

    # -- WURFL Updater allows seamless update of WURFL engine with new data downloaded from Scientiamobile.
    # -- Udater configuration must be done after wurfl_root.
    # -- WURFL file should be either .zip or .xml.gz and match wurfl_root file type.
    # -- Put your personal updater url taken from Scientiamobile customer Vault. Please note that in case your license is expired, 
    # -- NGINX won't start with updater configured.
    # -- Valid values for the updater checking frequency (how often the updater checks for any new WURFL data file
    # -- to be downloaded and used by the engine) are DAILY,WEEKLY.
    # -- Updater log file (wurfl-updater.log) may be found in "wurfl_root" folder. The folder and wurfl.zip file should be writable, 
    # -- and a wurfl.zip file must already be present in order for the Updater to determine whether or not it has to pull an update.
    # -- by NGINX process
    #wurfl_updater   https://data.scientiamobile.com/xxxxx/wurfl.zip DAILY;

    # -- WURFL patches definition (as much as needed, patches will be applied in the same order as specified in this conf file)
    #wurfl_patch     /path/to/patch1.xml;
    #wurfl_patch     /path/to/patch2.xml;
    #wurfl_patch     /path/to/patch3.xml;

    # Increase the variable hash size
    variables_hash_max_size 1024;
    variables_hash_bucket_size 1024;

    # -- WURFL cache: one of the following
    wurfl_cache_lru 100000;
    #wurfl_cache_null;

    # -- WURFL properties (formerly "WURFL default variables")
    # -- Since WURFL API version 1.8.0.0, WURFL default variables except "wurfl_id" are no longer injected by default
    # -- and have to be explicitly specified.
    #wurfl_request_property  wurfl_root_id;
    #wurfl_request_property  wurfl_isdevroot;
    #wurfl_request_property  wurfl_useragent;
    #wurfl_request_property  wurfl_info;
    #wurfl_request_property  wurfl_api_version;
    #wurfl_request_property  wurfl_last_load_time;
    #wurfl_request_property  wurfl_normalized_useragent;

    # -- WURFL user requested static capabilities (as an example, this is not a complete list)
    #wurfl_request_capability is_tablet;
    #wurfl_request_capability is_wireless_device;

    # -- WURFL user requested virtual capabilities (as an example, this is not a complete list).
    # -- Since WURFL API version 1.7.1.0, virtual capabilities are no longer injected by default
    # -- and have to be explicitly specified.
    #wurfl_request_capability advertised_device_os;
    #wurfl_request_capability is_android;

    # ...

    server {

        # ...

        # -- WURFL injection rules
        #
        # -- The following rule lists define which urls hasn't/has to be injected with wurfl data.
        #
        # -- The urls will be processed in this manner:
        #
        # -- Check if a wurfl_do_not_process_url regex matches the url. If yes, the url
        # -- itself will not be injected and no further checks will be made
        # -- If no wurfl_do_not_process_url match, check if a wurfl_process_url regex
        # -- matches the url. If yes, the url itself will be injected
        #
        # -- The fallback behaviour in case the url doesn't match any
        # -- wurfl_do_not_process_url/wurfl_process_url rule is:
        # -- INJECTION if the list of wurfl_process_url is empty
        # -- NO INJECTION if the list of wurfl_process_url is not empty
        #
        #
        # ---- Black list: defines which urls hasn't to be injected with wurfl data
        # ---- syntax: wurfl_do_not_process_url <url regex> <rule name>
        #
        # wurfl_do_not_process_url .*\.(gif|jpeg|png|css) "Static contents";
        # wurfl_do_not_process_url .*\/img\/.* "All in img folder;
        #
        #
        # ---- White list: defines which urls has to be injected with wurfl data
        # ---- syntax: wurfl_process_url <url regex> <rule name>
        #
        # wurfl_process_url .*\.(php|php4|php5) "Php scripts";
        # wurfl_process_url .*\.(jsp|asp) "Jsp/Asp scripts";

        # -- Trace wurfl injections (default value = off)
        # -- The log format is:
        # -- WURFL: Server <server name> - Whitelist hit - resource: <url> - regex: <url regex> (rule name:<rule name>)
        # -- WURFL: Server <server name> - Blacklist hit - resource: <url> - regex: <url regex> (rule name:<rule name>)
        # -- <server name> is the value of server_name property, or, if empty, a number indicating
        # -- which server section the log is relative to
        #
        # wurfl_log_header_injection on;

        # ...

        # Example of uses with PHP FastCGI (Note: PHP is not a requirement!)
        location ~ \.php$ {
            root           html;
            fastcgi_pass   127.0.0.1:9000;
            fastcgi_index  index.php;
            fastcgi_param  SCRIPT_FILENAME  /var/www$fastcgi_script_name;
            include        fastcgi_params;

            ############ WURFL data passed to fastcgi PHP scripts
            #
            ######## WURFL properties (formerly "WURFL default variables"): add here to make them available to fastcgi PHP scripts.
            ######## Property "wurfl_id" is injected by default (see http section) and could be specified here
            fastcgi_param  WURFL_ID                         $wurfl_id;
            #fastcgi_param  WURFL_ROOT_ID                    $wurfl_root_id;
            #fastcgi_param  WURFL_ISDEVROOT                  $wurfl_isdevroot;
            #fastcgi_param  WURFL_ORIGINAL_USERAGENT         $wurfl_useragent;
            #fastcgi_param  WURFL_INFO                       $wurfl_info;
            #fastcgi_param  WURFL_API_VERSION                $wurfl_api_version;
            #fastcgi_param  WURFL_LAST_LOAD_TIME             $wurfl_last_load_time;
            #fastcgi_param  WURFL_NORMALIZED_USERAGENT       $wurfl_normalized_useragent;
            #
            #
            ######## WURFL capabilities: headers injected by into Request by wurfl engine.
            ######## Add here to make them available to fastcgi PHP scripts.
            #
            #### specify the static capabilities needed in webapp (as an example, this is not a complete list)
            #fastcgi_param  WURFL_IS_TABLET                  $wurfl_cap_is_tablet;
            #fastcgi_param  WURFL_IS_WIRELESS_DEVICE         $wurfl_cap_is_wireless_device;
            #
            #### specify the virtual capabilities needed in webapp (as an example, this is not a complete list)
            #fastcgi_param  WURFL_ADVERTISED_DEVICE_OS       $wurfl_cap_advertised_device_os;
            #fastcgi_param  WURFL_IS_ANDROID                 $wurfl_cap_is_android;
        }
        # ...
    }
}

Important Note

If you are using NGINX with the builtin wurfl updater make sure that the folder that contains the wurfl.zip file is writable to all. Even if nginx starts as root, worker child processes might run as unpriviliged user ("nobody" by default) and this prevents workers engine updater to correctly update the wurfl.zip file during normal operations.

WURFL NGINX Module command table

The following table shows directives which can be applied to configure and use our WURFL NGINX Module.

Section Syntax Description Availability
http wurfl_enable Enable/Disable the WURFL module. Since version 1.8.4.1, the WURFL module is disabled by default and you will need to issue the wurfl_enable on; to enable it. Otherwise all WURFL commands will be ignored and the WURFL module will not be loaded.
Possible values: on/off. Default value: off.
1.9
http wurfl_root Defines the location (path) of the WURFL data file. 1.4
wurfl_updater Allows seamless update of WURFL engine with new data downloaded from Scientiamobile. This directive must follow wurfl_root.
It takes two parameters:
• the data url (taken from your personal Scientiamobile Vault account, choosing between two data file types: .zip or .xml.gz)
Take care that wurfl_root file type and wurfl_updater data url file types match so you may need to change the wurfl_root file type accordingly.
Please note that in case your license is expired, NGINX won't start with updater configured.
• the updater checking frequency (how often the updater checks for any new WURFL data file to be downloaded and used by the engine) which you can choose between DAILY and WEEKLY.
In order to let the Updater perform its activities both the wurfl_root folder and file must be writable by NGINX.
The wurfl-updater.log file in wurfl_root folder will contains details on Updater activity.
1.8.3
wurfl_patch Adds one or more custom patch files to the WURFL repository. 1.4
wurfl_target_default
or
wurfl_target_fast_desktop_browser_match
or
wurfl_target_performance
or
wurfl_target_accuracy
These configuration options are deprecated and will be removed in a future release. 1.4
wurfl_cache_lru
or
wurfl_cache_null
In order to increase performance while processing real HTTP traffic, we suggest setting up a LRU cache. The LRU caching strategy will speed up lookup operations on processed User Agents by keeping them in an LRU map. By default the cache will be set to 30000 entries which accounts for 7 to 10 MB of additional memory usage. Specific concerns regarding memory usage apart, users are advised to size their cache generously (100,000 or more) to increase performance. For more information, please see LRU Cache Mechanism. 1.4
wurfl_request_capability Enables injection of WURFL Static Capabilities/Virtual Capabilities in the HTTP request headers with your specified parameter(s).
Syntax: wurfl_request_capability <static capability/virtual capability name>
You can find the complete list of WURFL Static Capabilities and WURFL Virtual Capabilities here.
If a Static Capability/Virtual Capability is specified in a wurfl_request_capability command, WURFL NGINX module set a NGINX variable whose name is
wurfl_cap_
1.4
wurfl_useragent_priority_override_sideloaded_browser_useragent
or
wurfl_useragent_priority_use_plain_useragent
These configuration options are deprecated and will be removed in a future release. 1.5.2
wurfl_request_property Enables injection of WURFL Properties (see section WURFL Properties below) in the HTTP request headers.
Syntax: wurfl_request_property <variable name>
Not mandatory.
1.8
server wurfl_do_not_process_url A regular expression defining which urls hasn't to be injected with wurfl data. Syntax: wurfl_do_not_process_url <url regex> <rule name>
Not mandatory.
You can specify more than one wurfl_do_not_process_url
The urls will be processed in this manner:
Check if a wurfl_do_not_process_url regex matches the url.
• If yes, the url itself will not be injected and no further checks will be made.
• If no, check if a wurfl_process_url regex matches the url.
• • If yes, the url itself will be injected
The fallback behaviour in case the url doesn't match any wurfl_do_not_process_url/wurfl_process_url rule is:
• INJECTION if the list of wurfl_process_url is empty
• NO INJECTION if the list of wurfl_process_url is not empty
1.8
wurfl_process_url A regular expression defining which urls has to be injected with wurfl data.
Syntax: wurfl_process_url <url regex> <rule name>
Not mandatory.
You can specify more than one wurfl_process_url.
The urls will be processed in this manner:
Check if a wurfl_do_not_process_url regex matches the url.
• If yes, the url itself will not be injected and no further checks will be made.
• If no, check if a wurfl_process_url regex matches the url.
• • If yes, the url itself will be injected
The fallback behaviour in case the url doesn't match any wurfl_do_not_process_url/wurfl_process_url rule is:
• INJECTION if the list of wurfl_process_url is empty
• NO INJECTION if the list of wurfl_process_url is not empty
1.8
wurfl_log_header_injection Enable/disable logging of occurred injections.
You can choose between on/off
Not mandatory. (default value is off)
1.8

WURFL Properties

The WURFL NGINX module sets some useful convenience NGINX variables to retrieve information regarding the currently active WURFL configuration.
These variables are automatically calculated and are injected in HTTP requests if specified in a wurfl_request_property command.
Please note that the wurfl_id variable is injected by default so you don't have to sepcify it in a wurfl_request_property command.

WURFL Properties Table

Variable Name Contents Availability
wurfl_id Contains the device ID of the matched device. It is injected by default so you don't have to specify it in a wurfl_request_property command 1.4
wurfl_root_id Contains the device root ID of the matched device. 1.4
wurfl_isdevroot Tells if the matched device is a root device. Possible values are "TRUE" or "FALSE" 1.4
wurfl_useragent The original useragent coming with this particular web request 1.5.1
wurfl_api_version Contains a string representing the currently used Libwurfl API version 1.5.1
wurfl_engine_target This property is deprecated and will be removed in a future release. 1.5.1
wurfl_info A string containing information on the parsed WURFL data file and its full path 1.5.1
wurfl_last_load_time Contains the UNIX timestamp of the last time WURFL has been loaded successfully. 1.5.1
wurfl_normalized_useragent The normalized useragent. 1.5.1.3
wurfl_useragent_priority This property is deprecated and will be removed in a future release. 1.5.2

Running NGINX Web Server

Once you have configured NGINX and are ready to test the installion by launching the webserver, use the following command:

$ sudo /usr/local/nginx/sbin/nginx

Tip: This guide only instructs you on how to start and stop NGINX manually. There are many different online guides that shows how to build init scripts for NGINX, here is an example site.

To see if NGINX is running, go to your server's IP address. You should see a 'welcome to NGINX' message, indicating that NGINX is installed. Alternatively, check to see if the NGINX process is running by executing this command:

$ sudo ps aux | grep nginx

To stop the NGINX service manually, you can take advantage of the pid file to identify process ID.

$ sudo kill `cat /usr/local/nginx/logs/nginx.pid`

Verify WURFL Installation

At the end of the installation procedure, your NGINX web server will be enhanced with the new WURFL API capabilities to support a variety of use-cases that rely on your NGINX instance becoming aware of device information. Among other things, NGINX will augment HTTP requests with new headers such as: X-Wurfl-Is-Tablet, X-Is-Full-Desktop and the relative values as defined in WURFL. This will enable services deployed downstream to take advantage of device detection in the simplest way possible (getHeader() and similar).

Note: Pre-selected capabilities and header names are shown for this example (the actual choice of header names is configurable).

The WURFL Module for NGINX enables organizations to leverage device and browser information within the very tools, frameworks and programming languages that they have previously elected to use. For example, assuming that a RIM BlackBerry 8830 device requests a PHP phpinfo() page (assuming that PHP is installed), the following information would be returned by the system thanks to the WURFL NGINX Module:

Figure 1.

PHP variables now include HTTP headers that contain device information. This information can be leveraged by downstream applications to tailor the content to the capabilities of the requesting device. For example, assuming we choose the X-Wurfl-* notation for the environment variables or HTTP headers, the following PHP code can be used to determine if the requesting device is a tablet:

// PHP
if ($_SERVER['X-Wurfl-Is-Tablet'] == 'true') {
    //Do whatever makes sense for a tablet
}

Java, ASP.Net, Perl, Python and everything that can run in a CGI environment can take advantage of the functionality.

Running PHP Script on NGINX

In order to run the example script above, you will need to install PHP FastCGI. It is highly recommended you use PHP-FPM (FastCGI Process Manager) for its features and ease of installation. Use the following command to install the latest PHP-FPM package:

$ sudo apt-get install php5-fpm

Then run PHP5-FPM in the background, using the following init.d script to start the process:

$ sudo /etc/init.d/php-fpm start

At this point we are ready to create a simple test script. Test the capabilities filters with the following PHP script test.php to put in /usr/local/nginx/html:

//PHP - test.php
phpinfo(INFO_ENVIRONMENT);

The test.php script will return an extended result of what can be seen in Figure 1. You can use any kind of mobile device to load the site and see their unique device capabilities.


© 2024 ScientiaMobile Inc.
All Rights Reserved.

NOTICE: All information contained herein is, and remains the property of ScientiaMobile Incorporated and its suppliers, if any. The intellectual and technical concepts contained herein are proprietary to ScientiaMobile Incorporated and its suppliers and may be covered by U.S. and Foreign Patents, patents in process, and are protected by trade secret or copyright law. Dissemination of this information or reproduction of this material is strictly forbidden unless prior written permission is obtained from ScientiaMobile Incorporated.