GSP
Quick Navigator

Search Site

Unix VPS
A - Starter
B - Basic
C - Preferred
D - Commercial
MPS - Dedicated
Previous VPSs
* Sign Up! *

Support
Contact Us
Online Help
Handbooks
Domain Status
Man Pages

FAQ
Virtual Servers
Pricing
Billing
Technical

Network
Facilities
Connectivity
Topology Map

Miscellaneous
Server Agreement
Year 2038
Credits
 

USA Flag

 

 

Man Pages


Manual Reference Pages  -  NEUBOT (1)

NAME

neubot - The network neutrality bot \$1 \n[an-margin] level \n[rst2man-indent-level] level margin: \n[rst2man-indent\n[rst2man-indent-level]] - \n[rst2man-indent0] \n[rst2man-indent1] \n[rst2man-indent2]

CONTENTS

Synopsis
Description
Implemented Tests
Subcommands
Files
Examples
Web Interface Files
Web Api
Bittorrent Data Format
Raw Test Data Format
Speedtest Data Format
Data Processing Language
Privacy
Author
Copyright
See Also

SYNOPSIS

neubot --help
neubot -V
neubot subcommand ...

DESCRIPTION

 Neubot is a free-software Internet bot that gathers network performance data useful to investigate network neutrality. Neubot runs in the background and periodically performs active network tests with test servers provided by  Measurement Lab (a distributed server platform that hosts open network measurement tools). Implemented tests are described in the  IMPLEMENTED TESTS section.

Neubot does not perform any network test until you grant it the permission to collect and publish your IP address for research purposes. Neubot behaves like this because it is developed in the European Union; therefore, it must comply with European privacy laws (which consider IP addresses personal data). See the  PRIVACY section for more info.

Neubot is a background process. You can control it by using its subcommands, its web interface or its web API. Neubot listens for web requests at http://127.0.0.1:9774/. To access the web interface, use either your favorite browser or the viewer subcommand. To change the address and/or port where Neubot listens for web requests, edit /etc/neubot/api (see the  FILES section).

Neubot runs with the privileges of the _neubot user, or with the ones of the user indicated in the /etc/neubot/users configuration file (see the  FILES section).

The command line interface allows you to get the usage string (neubot --help), get the version number (neubot -V), and run a Neubot\(aqs subcommand (neubot subcommand...).

IMPLEMENTED TESTS

All Neubot tests receive and send random data. Neubot does not monitor the user\(aqs traffic.

Neubot implements three active network tests: bittorrent, raw and speedtest. For each test, there is a Neubot subcommand that allows to run the test immediately. Moreover, Neubot schedules one of the three tests at random every 23 - 27 minutes.

The bittorrent test emulates BitTorrent peer-wire protocol and estimates the round-trip time, the download and the upload goodput (i.e. the application-level speed measured at the receiver). It uses the time that connect() takes to complete as an estimator of the round-trip time. It estimates the goodput by dividing the amount of transferred bytes by the elapsed time. To avoid consuming too much user resources, the bittorrent test adapts the number of bytes to transfer such that the test runs for about five seconds.

The raw test performs a raw 10-second TCP download to estimate the download goodput. Also, it estimates the goodput from the round-trip time, and it collects statistics about the TCP sender by using Web100 (see http://www.web100.org). In addition, it estimates the round-trip time in two ways: (1) by measuring the time that connect() takes to complete (like bittorrent) and (2) by measuring the average time elapsed between sending a small request and receiving a small response.

The speedtest test emulates HTTP and estimates the round-trip time, the download and the upload goodput. It estimates the round-trip time in two ways: (1) by measuring the time that connect() takes to complete (like bittorrent) and (2) by measuring the average time elapsed between sending a small request and receiving a small response (like raw). It estimates the goodput by dividing the amount of transferred bytes by the elapsed time. To avoid consuming too much user resources, the bittorrent test adapts the number of bytes to transfer such that the test runs for about five seconds.

SUBCOMMANDS

This section documents Neubot\(aqs subcommands.
neubot bittorrent [-6fv] [-A address] [-p port]
  Asks Neubot to run a bittorrent test using the web API and fails if Neubot is not running in the background.

Accepts the following options:

-6 Prefer IPv6 to IPv4.
-A address
  Address of the remote test server.
-f Force the test. Run the test in the local process context (instead of using the web API) and override privacy settings if needed. Useful for developers and for debugging.
-p port Port of the remote test server.
-v Makes the command more verbose.

neubot database [-f database] [action]
  Performs the specified action or prints the database\(aqs path if no action is specified. We do not recommended to use this command to modify the database while Neubot is running, since Neubot does not expect the database to change while it is running, so it won\(aqt pick the changes up. This command requires root privileges to modify the database: if you are not root, the database is opened in readonly mode.

Accepts the following options:

-f database
  Force file. Forces the command to use database instead of the default database path.

Implements the following actions:
delete_all
  Deletes all the results in the database.
dump Dumps to the standard ouput the content of the database in JSON format.
prune Removes results older than one year.
regen_uuid
  Generates a new random unique identifier (or UUID) for Neubot. (See the  PRIVACY section for more on the UUID).
show Pretty prints to the standard ouput the content of the database in JSON format.
neubot privacy [-Pt] [-D setting=value] [-f database]
  Manage privacy settings. When invoked without options, this command prints the current privacy settings values.

Accepts the following options:

-D setting=value
  Turn on (nonzero) and off (zero) the specified privacy setting.

This command just modifies the database: you have to restart Neubot to make changes effective. To modify privacy settings when Neubot is running, we recommend to use the web interface.

Privacy settings:

informed
  The user has read the privacy policy.
can_collect
  The user provides the permission to collect his/her IP address for research purposes.
can_publish
  The user provides the permission to publish his/her IP address allowing anyone to reuse it for research purposes.
-f database
  Force file. Forces the command to use database instead of the default database path.
-P Prints privacy policy on the standard output.
-t Test. Exits with success (exit value 0) if all privacy settings all nonzero. Exits with failure (exit value nonzero) if at least one setting is zero.
neubot raw [-6fv] [-A address] [-p port]
  Asks Neubot to run a raw test using the web API and fails if Neubot is not running in the background.

Accepts the following options:

-6 Prefer IPv6 to IPv4.
-A address
  Address of the remote test server.
-f Force the test. Run the test in the local process context (instead of using the web API) and override privacy settings if needed. Useful for developers and for debugging.
-p port Port of the remote test server.
-v Makes the command more verbose.
neubot speedtest [-6fv] [-A address] [-p port]
  Asks Neubot to run a speedtest test using the web API and fails if Neubot is not running in the background.

Accepts the following options:

-6 Prefer IPv6 to IPv4.
-A address
  Address of the remote test server.
-f Force the test. Run the test in the local process context (instead of using the web API) and override privacy settings if needed. Useful for developers and for debugging.
-p port Port of the remote test server.
-v Makes the command more verbose.
neubot start
  On MacOS this command runs launchctl(1), which in turn starts Neubot. You must be root to run this command. On MacOS, Neubot\(aqs installer configures the system to launch Neubot at startup; i.e. you do not typically need to run this command.

(On MacOS, Neubot is implemented by two daemons: the usual unprivileged daemon and a privileged daemon. The latter controls the former and periodically forks an unprivileged child to check for updates.)

On MacOS, the start command accepts the following options:

-a Auto-updates. When this flag is specified, the privileged daemon periodically forks an unprivileged child to check for updates.
-d Debug. When this flag is specified, Neubot runs in the foreground.
-v Verbose. When this flag is specified, the start command is verbose (i.e. it prints on the standard error the commands it is about to invoke).

When both -v and -d are specified, Neubot runs in verbose mode in the foreground.

At boot time, launchctl(1) starts Neubot with the -a and -d command line options.

On other UNIX systems, the start command forks the Neubot daemon, which drops root privileges and runs in the background. On such systems, this command does not accept any command line option.
neubot status
  This command asks the status of Neubot using the web API. It returns 0 if connect() succeeds and the response is OK, nonzero otherwise.

On MacOS this command accepts the -v option, which makes it more verbose. On other UNIX systems, it does not accept any command line option.

neubot stop
  On MacOS, this command runs launchctl(1), which in turn stops Neubot. You must be root to run this command. On MacOS, this command accepts the -v option, which makes it more verbose.

On other UNIX systems, this command uses the web API to request Neubot to exit.

neubot viewer
  This command shows the web interface by embedding a web rendering engine into a window manager\(aqs window. Currently, the only implemented viewer is based on python-webkit and pygtk.

FILES

Assuming that Neubot is installed at /usr/local, this is the list of the files installed.
/etc/neubot/api
  Configuration file that indicates the endpoint where Neubot should listen for web API requests. Example (which also shows the syntax and indicates the default values):


#
# /etc/neubot/api - controls address, port where Neubot listens
# for incoming web API requests.
#
address 127.0.0.1  # Address where the listen
port 9774          # Port where to listen


/etc/neubot/users
  Configuration file that indicates the unprivileged user names that Neubot should use. Example (which also shows the syntax and indicates the default values):


#
# /etc/neubot/users - controls the unprivileged user names used
# by Neubot to perform various tasks.
#
update_user _neubot_update  # For auto-updates (MacOS-only)
unpriv_user _neubot         # For network tests


/usr/local/bin/neubot
  The Neubot executable script.
/usr/local/share/neubot/
  Location where Neubot Python modules are installed.
/usr/local/share/neubot/www/
  Location where the web interface files are installed. The web interface is described in the  WEB INTERFACE FILES section.
/var/lib/neubot/database.sqlite3
  System-wide results database for Linux systems, created when Neubot starts for the first time.
/var/neubot/database.sqlite3
  System-wide results database for non-Linux systems, created when Neubot starts for the first time.

EXAMPLES

In this section, we represent the unprivileged user prompt with $ and the root user prompt with #.

Run on-demand bittorrent test:


$ neubot bittorrent


Run on-demand raw test:


$ neubot raw


Run on-demand speedtest test:


$ neubot speedtest


Start Neubot:


# neubot start


Stop Neubot:


# neubot stop  # MacOS
$ neubot stop  # other UNIX


Run Neubot in the foreground with verbose logging:


# neubot start -dv                       # MacOS
$ neubot agent -v -D agent.daemonize=no  # other UNIX


Export Neubot results to JSON:


# neubot database dump > output.json


Read Neubot\(aqs privacy policy:


$ neubot privacy -P


Run Neubot command from the sources directory:


$ ./UNIX/bin/neubot command


WEB INTERFACE FILES

Here we provide a brief description of the core files of the web interface:
css/ Directory that contains CSS files.
favicon.ico
  Neubot\(aqs favicon.
footer.html
  Common footer for all web pages (Neubot uses server-side includes).
header.html
  Common header for all web pages (Neubot uses server-side includes).
img/ Directory that contains images.
js/ Directory that contains javascript files. In addition to jQuery and jqPlot, it contains the following scripts:
js/contrib.js
  Helper functions from many authors.
js/i18n.js
  Implementation of web user interface internationalization (aka i18n).
js/index.js
  Contains functions to retrieve and process the state of Neubot.
js/log.js
  Contains code to retrieve and process Neubot logs.
js/privacy.js
  Contains code to query and modify privacy settings.
js/results.js
  Contains code to process Neubot results, as well as code to display them as plots and tables.
js/settings.js
  Contains code to retrieve and modify Neubot settings.
js/state.js
  Helper code for retrieving and processing Neubot state.
js/update.js
  Minimal script included by updater.html. It just sets the active tab in the web interface.
js/utils.js
  Miscellaneous helper functions.

lang/ Directory that contains one javascript file for each language in which the web interface is translated. Each of these javascripts contains a dictionary, named LANG, that maps a string (or a key representing a string) to its translation.

In javascript, you mark strings for translation by wrapping them with i18n.get() calls. For example, to indicate that the string "Disable automatic tests" should be translated, you should write:


...
i18n.get("Disable automatic tests");


In HTML code, you mark the content of an HTML tag for translation by adding the tag to the i18n class. Differently from javascript, we don\(aqt map the content of an HTML tag to its translation; instead, we map a key that represents the HTML tag content to its translation. The key is another HTML class, which must start with i18n_, as in the following example:


...
<p class="i18n i18n_foobar">Neubot web interface</p>


To translate the two examples above in, for example, Italian you edit the www/lang/it.css file and add:


var LANG = {
    ...
    "Disable automatic tests": "Disabilita test automatici",
    "i18n_foobar": "Interfaccia web di Neubot",
    ...
};


log.html
  Shows Neubot logs.
not_running.html
  Page displayed when Neubot is not running.
privacy.html
  Shows, and allows to modify, privacy settings.
results.html
  The results page, dynamically filled by javascript using Neubot web API. It allows you to see the results of recent experiments, both in form of plots and tables.
settings.html
  Shows (and allows to modify) Neubot settings.
test/ Directory that contains a foo.html and a foo.json file for each test foo. The list of available tests in results.html is automatically generated from the files in this directory.
test/foo.html
  Description of the foo test. It is included into the results.html page when the test is selected.
test/foo.json
  Description of the plots and tables included into results.html when test foo is selected. The format of the JSON is documented into the  WEB API section of this manual page.
test/foo.json.local
  When foo.json.local exists, Neubot will use it (instead of foo.json) to prepare plots and tables in results.html. Allows the user to heavily customize the results page for test foo.
update.html
  Page displayed on Windows when Neubot needs to be manually updated. Now that automatic updates are implemented, it should never pop up.

WEB API

To access Neubot API, you send HTTP requests to the address and port where Neubot is listening (which is 127.0.0.1:9774 by default, and which can be changed by editing /etc/neubot/api).

Here is a detailed description of each API.
/api This API is an alias for /api/.
/api/ This API allows you to get (GET) the list of available APIs, encoded as a JSON.

Returned JSON example:


 [
  "/api",
  "/api/",
  "/api/config",
  "/api/data",
  "/api/debug",
  "/api/exit",
  "/api/index",
  "/api/log",
  "/api/results",
  "/api/runner",
  "/api/state",
  "/api/version"
]


/api/config[?options]
  This API allows to you get (GET) and set (POST) the variables that modify the behavior of Neubot.

GET returns a dictionary, encoded using JSON, that maps each variable to its value. POST sends a url-encoded string, which contains one or more variable=new_value atoms separated by &.

The API accepts the following query-string options:

debug=integer [default: 0]
  When nonzero, the API returns a pretty-printed JSON. Otherwise, the JSON is serialized on a single line.
labels=integer [default: 0]
  When nonzero, returns the description of the variables instead of their values.

Returned JSON example:


{
 "enabled": 1,
 "negotiate.max_thresh": 64,
 "negotiate.min_thresh": 32,
 "negotiate.parallelism": 7,
 "privacy.can_collect": 1,
 "privacy.can_publish": 1,
 "privacy.can_informed": 1,
 ...
 "uuid": "0964312e-f451-4579-9984-3954dcfdeb42",
 "version": "4.2",
 "www.lang": "default"
}


We have not standardized variable names yet. Therefore, we don\(aqt provide here a list of variable names, types, and default values.
/api/data?test=string[&options]
  This API allows you to retrieve (GET) the data collected during Neubot tests. As we have a single API for all tests, you must provide the test name using the query string.

This API returns a JSON that serializes a list of dictionaries, in which each dictionary is the data collected during a test. We dedicate a section of the manual page to the structure returned by each test.

This API accepts the following query-string parameters:

debug=integer [default: 0]
  When nonzero, the API returns a pretty-printed JSON. Otherwise, the JSON is serialized on a single line.
since=integer [default: 0]
  Returns only the data collected after the specified time (indicated as the number of seconds elapsed since midnight of January, 1st 1970).
test=string
  This parameter is mandatory and specifies the test whose data you want to retrieve.
until=integer [default: 0]
  Returns only the data collected before the specified time (indicated as the number of seconds elapsed since midnight of January, 1st 1970).
/api/debug
  This API allows you to get (GET) text/plain information about Neubot internals, which is typically useful for debugging purposes. As such, the consistency of the output format is not guaranteed.

Returned text example:


{\(aqWWW\(aq: \(aq/usr/share/neubot/www\(aq,
 \(aqnotifier\(aq: {\(aq_subscribers\(aq: {},
           \(aq_timestamps\(aq: {\(aqstatechange\(aq: 1336727245277393,
                           \(aqtestdone\(aq: 1336727245277246}},
 \(aqqueue_history\(aq: [],
 \(aqtypestats\(aq: {\(aqABCMeta\(aq: 26,
               \(aqBackendNeubot\(aq: 1,
               \(aqBackendProxy\(aq: 1,
               ...
              }}


/api/exit
  When this API is invoked, Neubot exits immediately (i.e. without sending any response).

Don\(aqt use this API to shut down Neubot on MacOS, use the neubot stop command instead. This API, in fact, has effect on the unprivileged Neubot process only, and the privileged process will respawn the unprivileged process once it notices it died.

/api/index
  This API uses 302 Found and Location to redirect the caller to either index.html (if privacy settings are OK) or on privacy.html (if privacy settings are not OK).
/api/log[?options]
  This API allows you to get (GET) Neubot logs, as a list of dictionaries. Each dictionary represents a log record and contains the following fields:
timestamp (integer)
  Time when this log was generated, expressed as number of seconds elapsed since midnight of January, 1st 1970.
severity (string)
  The log message severity; one of: DEBUG, INFO, WARNING, and ERROR.
message (string)
  The log message string.

This API accepts the following query-string options:
debug (int) [default: 0]
  If nonzero, the API formats logs like they are printed on the system logger (i.e. as a text/plain sequence of lines). Otherwise, the API returns the JSON list of dictionaries described above.
reversed (int) [default: 0]
  If nonzero logs are reversed (i.e. the most recent log record is the first element of the list). Otherwise logs are returned in natural order (the most recent log record is the last element of the list).
verbosity (int) [default: 1]
  When the verbosity is less than 1, only ERROR and WARNING messages are returned. When the verbosity is 1, the API returns also INFO messages. When the verbosity is greater than 1, the API returns also INFO and DEBUG messages.

Returned JSON example:


[
 {
  "message": "raw_negotiate: not reached final state",
  "severity": "WARNING",
  "timestamp": 1366195042
 },
 {
  "message": "raw_negotiate: bad response",
  "severity": "ERROR",
  "timestamp": 1366236483
 },
 {
  "message": "raw_negotiate: not reached final state",
  "severity": "WARNING",
  "timestamp": 1366236484
 }
]


/api/results?test=string[&options]
  This API allows the web interface to get (GET) information on how to format results. It returns a dictionary, encoded as JSON, that indicates the plots and the tables to be generated in the results.html page for the selected test (which is either the test specified via query string or the default test, speedtest, if none was specified).

The dictionary for test foo is generated using www/test/foo.json (or www/test/foo.json.local) as template and contains the following fields:

available_tests (list of strings)
  List that contains the name of all the available tests.
description (string)
  String that contains a long description of the selected test. This is the content of www/test/foo.html.
plots (list of dictionaries)
  List of dictionaries. Each dictionary contains the instructions to generate a plot:
datasets (list of dictionaries)
  List of dictionaries. Each dictionary contains the instructions to plot one serie of data:
label (string)
  Label to use in the legend.
marker (string)
  Indicates the marker to use, either circle or square.
recipe (list)
  LISP-like code that describes how to generate one point on the Y axis from one row of the selected test\(aqs data. We describe this lisp-like language in the  DATA PROCESSING LANGUAGE section of this manual page.
title (string)
  Title of the plot.
xlabel (string)
  Label for the X axis.
ylabel (string)
  Label for the Y axis.
selected_test
  The selected test name.
table (list of dictionaries)
  List of dictionaries. Each dictionary is one column of the table to be added to results.html:
label (string)
  Label of the column header.
recipe (list)
  LISP-like code that describes how to generate the value of the current column in the table from one row of the selected test\(aqs data. We describe this lisp-like language in the  DATA PROCESSING LANGUAGE section of this manual page.
title (string)
  The title of the test (e.g. \(aqBitTorrent test\(aq).
www_no_description (integer)
  Whether to include a description of the test in the results page (zero) or not (nonzero).
www_no_legend (integer)
  Whether to include a legend in the plots (zero) or not (nonzero).
www_no_plot (integer)
  Whether to generate plots (zero) or not (nonzero).
www_no_split_by_ip (integer)
  Whether to split the selected test\(aqs data by IP and plot a different line for each IP (zero) or not (nonzero).
www_no_table (integer)
  Whether to generate a table that contains the selected test\(aqs data (zero) or not (nonzero).
www_no_title (integer)
  Whether to include the title of the test in the results page (zero) or not (nonzero).

The API accepts the following query-string options:
debug=integer [default: 0]
  When nonzero, the API returns a pretty-printed JSON. Otherwise, the JSON is serialized on a single line.
test=string
  This parameter is mandatory and specifies the selected test.

Returned JSON example:


{
 "available_tests": [
     "raw",
     "speedtest",
     "bittorrent"
 ],
 "description": "...",
 "www_no_split_by_ip": 0,
 "title": "Your recent Speedtest results",
 "www_no_legend": 0,
 "selected_test": "speedtest",
 "www_no_plot": 0,
 "www_no_table": 0,
 "table": [
     {
         "recipe": ["to-datetime",
                     ["select", "timestamp", "result"]],
         "label": "Timestamp"
     },
     {
         "recipe": ["select", "internal_address", "result"],
         "label": "Internal address"
     },
     {
         "recipe": ["select", "real_address", "result"],
         "label": "Real address"
     },
     {
         "recipe": ["select", "remote_address", "result"],
         "label": "Remote address"
     },
     {
         "recipe": ["to-millisecond-string",
                     ["select", "connect_time", "result"]],
         "label": "Connect time"
     },
     {
         "recipe": ["to-millisecond-string",
                     ["select", "latency", "result"]],
         "label": "Appl. latency"
     },
     {
         "recipe": ["to-speed-string",
                     ["select", "download_speed", "result"]],
         "label": "Download speed"
     },
     {
         "recipe": ["to-speed-string",
                     ["select", "upload_speed", "result"]],
         "label": "Upload speed"
     }
 ],
 "www_no_description": 0,
 "plots": [
     {
         "datasets": [
             {
                 "marker": "circle",
                 "recipe": ["to-speed",
                             ["select", "download_speed",
                              "result"]],
                 "label": "Dload"
             },
             {
                 "marker": "square",
                 "recipe": ["to-speed",
                             ["select", "upload_speed",
                              "result"]],
                 "label": "Upload"
             }
         ],
         "ylabel": "Goodput (Mbit/s)",
         "xlabel": "Date",
         "title": "Download and upload speed"
     },
     {
         "datasets": [
             {
                 "marker": "circle",
                 "recipe": ["to-millisecond",
                             ["select", "latency", "result"]],
                 "label": "Appl. latency"
             },
             {
                 "marker": "square",
                 "recipe": ["to-millisecond",
                             ["select", "connect_time",
                              "result"]],
                 "label": "Connect time"
             }
         ],
         "ylabel": "Delay (ms)",
         "xlabel": "Date",
         "title": "Connect time and latency"
     }
 ],
 "www_no_title": 0
}


/api/runner?test=string[&options]
  This API allows the caller to schedule a test for immediate execution. If a test is already running the API returns an error 500, otherwise it returns 200.

The API accepts the following query-string options:

test=string
  This option is mandatory and indicates the name of the test that Neubot should schedule for execution.
streaming=integer [default: 0]
  When nonzero, Neubot streams logs generated during the test in the response body and closes the connection when the test is complete. Otherwise, the response body is an empty dictionary.

When you invoke tests from the command line (e.g. neubot bittorrent), streaming is the feature that allows to print logs generated by the test on the console.

Returned JSON example:


{}


Returned text example:


1366299354 [INFO] runner_core: Need to auto-discover first...
1366299355 [INFO] runner_mlabns: server discovery...
1366299356 [INFO] runner_mlabns: server discovery... done
1366299356 [INFO] raw_clnt: connection established with ...
1366299356 [INFO] raw_clnt: connect_time: 13.6 ms
1366299357 [INFO] raw_clnt: sending auth to server...
1366299357 [INFO] raw_clnt: sending auth to server... done
1366299357 [INFO] raw_clnt: receiving auth from server...
1366299357 [INFO] raw_clnt: receiving auth from server... done
1366299357 [INFO] raw_clnt: estimating ALRTT...
1366299357 [INFO] raw_clnt: alrtt_avg: 14.3 ms
1366299357 [INFO] raw_clnt: estimating ALRTT... done
1366299357 [INFO] raw_clnt: raw goodput test...
1366299367 [INFO] raw_clnt: raw goodput test... done
1366299367 [INFO] raw_clnt: goodput: 65.5 Mbit/s


/api/state[?options]
  This API allows you to get (GET) and track (via comet) the state of Neubot. The API returns a dictionary with the following fields:
current=string
  The name of the current state; one of: idle, rendezvous, negotiate, test, and collect.
events=dictionary
  A dictionary that maps the name of an event (a string) to the most recent value related to such event (a string, an integer, a list, or a dictionary).

While running, Neubot generates a limited set of events, which drive the web interface. For example, the test_download event value is used to update the download speed in the right sidebar of the web interface.

The list of generated events is not standardized yet, so we don\(aqt provide it here.

t=integer
  The identifier of the current event.

The API accepts the following query-string options:
debug=integer [default: 0]
  When nonzero, the API returns a pretty-printed JSON. Otherwise, the JSON is serialized on a single line.
t=integer
  When this option is present, Neubot does not return a response until the next event after the one identified by integer is fired (or until a timeout expires). This behavior allows to implement the comet pattern and to timely update the web interface with low overhead.
/api/version
  This API allows you to get (GET) the version number of Neubot, in text/plain format.

Returned text example:


0.4.15.7


BITTORRENT DATA FORMAT

We represent the data collected by the bittorrent test with a dictionary that contains the following fields:
connect_time (float)
  RTT estimated by measuring the time that connect() takes to complete, measured in seconds.
download_speed (float)
  Download speed measured by dividing the number of received bytes by the elapsed download time, measured in bytes per second.
internal_address (string)
  Neubot\(aqs IP address, as seen by Neubot. It is typically either an IPv4 or an IPv6 address.
neubot_version (string)
  Neubot version number, encoded as a floating point number and printed into a string. Given a version number in the format <major>.<minor>.<patch>.<revision>, the encoding is as follows:


<major> + 1e-03 * <minor> + 1e-06 * <patch>
        + 1e-09 * <revision>


For example, the 0.4.15.3 version number is encoded as 0.004015003.
platform (string)
  The operating system platform, e.g. linux2, win32.
privacy_can_collect (integer)
  The value of the can_collect privacy setting.
privacy_can_publish (integer)
  The value of the can_publish privacy setting.
privacy_informed (integer)
  The value of the informed privacy setting.
real_address (string)
  Neubot\(aqs IP address, as seen by the server. It is typically either an IPv4 or an IPv6 address.
remote_address (string)
  The server\(aqs IP address. It is typically either an IPv4 or an IPv6 address.
timestamp (integer)
  Time when the test was performed, expressed as number of seconds elapsed since midnight of January, 1st 1970.
upload_speed (float)
  Upload speed measured by dividing the number of sent bytes by the elapsed upload time, measured in bytes per second.
uuid (string)
  Random unique identifier of the Neubot instance, useful to perform time series analysis.

Example:


[
 {
  "connect_time": 0.003387928009033203,
  "download_speed": 4242563.145733707,
  "internal_address": "130.192.91.231",
  "neubot_version": "0.004015007",
  "platform": "linux2",
  "privacy_can_collect": 1,
  "privacy_can_publish": 1,
  "privacy_informed": 1,
  "real_address": "130.192.91.231",
  "remote_address": "194.116.85.224",
  "test_version": 1,
  "timestamp": 1366045628,
  "upload_speed": 4231443.875881268,
  "uuid": "7528d674-25f0-4ac4-aff6-46f446034d81"
 },
 ...


RAW TEST DATA FORMAT

We represent the data collected by the raw test with a dictionary that contains the following fields:
connect_time (float)
  RTT estimated by measuring the time that connect() takes to complete, measured in seconds.
download_speed (float)
  Download speed measured by dividing the number of received bytes by the elapsed download time, measured in bytes per second.
json_data (string)
  This string contains the serialization of a JSON object, which contains all the data collected during the test, both on the server and on the client side. The dictionary that we are describing, in fact, contains just a subset of the collected results. We can not store the full JSON object directly until Neubot\(aqs database module and web interface are ready to process it.
internal_address (string)
  Neubot\(aqs IP address, as seen by Neubot. It is typically either an IPv4 or an IPv6 address.
latency (float)
  RTT estimated by measuring the average time elapsed between sending a small request and receiving a small response, measured in seconds.
neubot_version (float)
  Neubot version number, encoded as a floating point number and printed into a string. Given a version number in the format <major>.<minor>.<patch>.<revision>, the encoding is as follows:


<major> + 1e-03 * <minor> + 1e-06 * <patch>
        + 1e-09 * <revision>


For example, the 0.4.15.3 version number is encoded as 0.004015003.
platform (string)
  The operating system platform, e.g. linux2, win32.
real_address (string)
  Neubot\(aqs IP address, as seen by the server. It is typically either an IPv4 or an IPv6 address.
remote_address (string)
  The server\(aqs IP address. It is typically either an IPv4 or an IPv6 address.
timestamp (integer)
  Time when the test was performed, expressed as number of seconds elapsed since midnight of January, 1st 1970.
uuid (string)
  Random unique identifier of the Neubot instance, useful to perform time series analysis.

Example:


[
 {
  "connect_time": 0.2981860637664795,
  "download_speed": 3607.120929707688,
  "internal_address": "130.192.91.231",
  "json_data": "...",
  "latency": 0.29875500202178956,
  "neubot_version": "0.004015007",
  "platform": "linux2",
  "real_address": "130.192.91.231",
  "remote_address": "203.178.130.237",
  "timestamp": 1365071100,
  "uuid": "7528d674-25f0-4ac4-aff6-46f446034d81"
 },
 ...


Once unserialized, the JSON object saved into the json_data field of the raw dictionary (henceforth, \(aqouter dictionary\(aq) is a dictionary that contains the following fields:
client (dictionary)
  A dictionary that contains data collected on the client side.
server (dictionary)
  A dictionary that contains data collected on the server side.

The client dictionary contains the following fields:
al_capacity (float)
  Median bottleneck capacity computed at application level (experimental).
al_mss (float)
  MSS according to the application level (information gathered using setsockopt(2)).
al_rexmits (list)
  Likely retransmission events computed at application level (experimental).
alrtt_avg (float)
  Same as latency in the outer dictionary.
alrtt_list (list of tuples)
  List of RTT samples estimated by measuring the average time elapsed between sending a small request and receiving a small response, measured in seconds.
connect_time (float)
  Same as connect_time in the outer dictionary.
goodput (dictionary)
  The dictionary contains the following fields:
bytesdiff
  Total number of received bytes.
ticks (float)
  Timestamp when this piece of data was collected, expressed as number of seconds elapsed since midnight of January, 1st 1970.
timediff (float)
  Total download time.
goodput_snap (list of dictionaries)
  List that contains a dictionary, which is updated roughly every second during the download, and which contains the following fields:
ticks (float)
  Time when the current dictionary was saved, expressed as number of seconds since midnight of January, 1st 1970.
bytesdiff (integer)
  Number of bytes received since stats were previously saved.
timediff (float)
  Number of seconds elapsed since stats were previously saved.
utimediff (float)
  Difference between current tms_utime field of the tms struct modified by times(3) and the previous value of the same field.
stimediff (float)
  Difference between current tms_stime field of the tms struct modified by times(3) and the previous value of the same field.
myname (string)
  Neubot\(aqs address (according to the server). This is same as real_address in the outer dictionary.
peername (string)
  Servers\(aqs address. This is same as server_address in the outer dictionary.
platform (string)
  Same as platform in the outer dictionary.
uuid (string)
  Same as uuid in the outer dictionary.
version (string)
  Same as neubot_version in the outer dictionary.

The server dictionary contains the following fields:
goodput (dictionary)
  The dictionary contains the following fields:
bytesdiff
  Total number of sent bytes.
ticks (float)
  Timestamp when this piece of data was collected, expressed as number of seconds elapsed since midnight of January, 1st 1970.
timediff (float)
  Total upload time.
goodput_snap (list of dictionaries)
  List that contains a dictionary, which is updated roughly every second during the upload, and which contains the following fields:
ticks (float)
  Time when the current dictionary was saved, expressed as number of seconds since midnight of January, 1st 1970.
bytesdiff (integer)
  Number of bytes sent since stats were previously saved.
timediff (float)
  Number of seconds elapsed since stats were previously saved.
utimediff (float)
  Difference between current tms_utime field of the tms struct modified by times(3) and the previous value of the same field.
stimediff (float)
  Difference between current tms_stime field of the tms struct modified by times(3) and the previous value of the same field.
myname (string)
  Servers\(aqs address. This is same as server_address in the outer dictionary.
peername (string)
  Neubot\(aqs address (according to the server). This is same as real_address in the outer dictionary.
platform (string)
  Same as platform in the outer dictionary.
timestamp (integer)
  Time when the server dictionary was created, expressed as number of seconds elapsed since midnight of January, 1st 1970.
version (string)
  Same as neubot_version in the outer dictionary.
web100_snap (list)
  A list that contains dictionaries. Each dictionary is a snapshot of the Web100 TCP state. We take one Web100 snapshot every second during the upload.

On the client side, this field is empty. We are working to identify the most interesting fields that is interesting to save.

Example:


[
 {
  "client": {
   "al_mss": 1448,
   "uuid": "7528d674-25f0-4ac4-aff6-46f446034d81",
   "goodput": {
    "bytesdiff": 128200,
    "timediff": 35.540810108184814,
    "ticks": 1365071098.203412
   },
   "al_rexmits": [],
   "connect_time": 0.2981860637664795,
   "alrtt_list": [
    0.31011295318603516,
    0.30966901779174805,
    0.29677391052246094,
    0.2957899570465088,
    0.29570794105529785,
    0.2956199645996094,
    0.29558706283569336,
    0.2956211566925049,
    0.2958400249481201,
    0.296828031539917
   ],
   "myname": "130.192.91.231",
   "peername": "203.178.130.237",
   "platform": "linux2",
   "version": "0.004015007",
   "al_capacity": 10982553.692585895,
   "alrtt_avg": 0.29875500202178956,
   "goodput_snap": [
    {
     "bytesdiff": 24616,
     "timediff": 1.0001380443572998,
     "ticks": 1365071063.66274,
     "stimediff": 0.0,
     "utimediff": 0.0
    },
    ...
   ]
  },
  "server": {
   "timestamp": 1365070933,
   "myname": "203.178.130.237",
   "peername": "130.192.91.231",
   "platform": "linux2",
   "version": "0.004015007",
   "goodput": {
    "bytesdiff": 131092,
    "timediff": 34.94503116607666,
    "ticks": 1365070933.95337
   },
   "goodput_snap": [
    {
     "bytesdiff": 31856,
     "timediff": 1.0005459785461426,
     "ticks": 1365070900.008885,
     "stimediff": 0.0,
     "utimediff": 0.0
    },
    ...
   ],
   "web100_snap": []
  }
 }


SPEEDTEST DATA FORMAT

We represent the data collected by the speedtest test with a dictionary that contains the following fields:
connect_time (float)
  RTT estimated by measuring the time that connect() takes to complete, measured in seconds.
download_speed (float)
  Download speed measured by dividing the number of received bytes by the elapsed download time, measured in bytes per second.
internal_address (string)
  Neubot\(aqs IP address, as seen by Neubot. It is typically either an IPv4 or an IPv6 address.
latency (float)
  RTT estimated by measuring the average time elapsed between sending a small request and receiving a small response, measured in seconds.
neubot_version (string)
  Neubot version number, encoded as a floating point number and printed into a string. Given a version number in the format <major>.<minor>.<patch>.<revision>, the encoding is as follows:


<major> + 1e-03 * <minor> + 1e-06 * <patch>
        + 1e-09 * <revision>


For example, the 0.4.15.3 version number is encoded as 0.004015003.
platform (string)
  The operating system platform, e.g. linux2, win32.
privacy_can_collect (integer)
  The value of the can_collect privacy setting.
privacy_can_publish (integer)
  The value of the can_publish privacy setting.
privacy_informed (integer)
  The value of the informed privacy setting.
real_address (string)
  Neubot\(aqs IP address, as seen by the server. It is typically either an IPv4 or an IPv6 address.
remote_address (string)
  The server\(aqs IP address. It is typically either an IPv4 or an IPv6 address.
timestamp (integer)
  Time when the test was performed, expressed as number of seconds elapsed since midnight of January, 1st 1970.
upload_speed (float)
  Upload speed measured by dividing the number of sent bytes by the elapsed upload time, measured in bytes per second.
uuid (string)
  Random unique identifier of the Neubot instance, useful to perform time series analysis.

Example:


[
 {
  "connect_time": 0.0017991065979003906,
  "download_speed": 11626941.501993284,
  "internal_address": "130.192.91.231",
  "latency": 0.003973397341641513,
  "neubot_version": "0.004015007",
  "platform": "linux2",
  "privacy_can_collect": 1,
  "privacy_can_publish": 1,
  "privacy_informed": 1,
  "real_address": "130.192.91.231",
  "remote_address": "194.116.85.237",
  "test_version": 1,
  "timestamp": 1365074302,
  "upload_speed": 10974865.674026133,
  "uuid": "7528d674-25f0-4ac4-aff6-46f446034d81"
 },
 ...


DATA PROCESSING LANGUAGE

The data processing language is a simple LISP-like language. As such, it describes processes whose goal is to transform pieces of collected data by using lists.

Differently from traditional LISP syntax, however, the data processing language is encoded using JSON.

The language implements the following operations:
[ divide , atom-or-list, atom-or-list]
  Divides the left atom (or list) by the right atom (or list) and returns the result.
[ map-select , atom, list]
  Cycles over the list and, for each element, it selects the field indicated by the atom.
[ parse-json , atom-or-list]
  Parses the value of the atom (or list) into an object.
[ reduce-avg , list]
  Computes the average value of the list.
[ select , atom, object]
  Selects the element of object indicated by atom.
[ to-datetime , atom-or-list]
  Converts atom (or list) to datetime string.
[ to-millisecond , atom-or-list]
  Converts atom (or list) to millisecond.
[ to-millisecond-string , atom-or-list]
  Converts atom (or list) to millisecond string.
[ to-speed , atom-or-list]
  Converts atom (or list) to speed (in bits per second).
[ to-speed-string , atom-or-list]
  Converts atom (or list) to speed string (in bits per second).
result
  The current piece of data we are processing.

Example (select the json_data field of the result, convert it to json, take the client field, take and compute the average of the alrtt_list field, convert the result to millisecond):


["to-millisecond",
  ["reduce-avg",
    ["select", "alrtt_list",
      ["select", "client",
        ["parse-json",
          ["select", "json_data", "result"]]]]]]


PRIVACY

Neubot collects your IP address, which is personal data according to European privacy laws. For this reason, Neubot needs to obtain your permission to collect your IP address for research purposes, as well as to publish it on the web for the same purpose. In addition, it also needs that you assert that you have read the privacy policy.

Without the assertion that you have read the privacy policy and the permission to collect and publish your IP address, Neubot can not perform automatic (or manual) tests.

You can read Neubot\(aqs privacy policy by running the neubot privacy -P command. The privacy policy is also available at:


http://127.0.0.1:9774/privacy.html


Of course, if you modified the address and/or port where Neubot listens, you need to update the URI accordingly.

In addition to the above, each Neubot is identified by a random unique identifier (UUID) that is used to perform time series analysis. We believe that this identifier does not brach your privacy: in the worst case, we would be able to say that a given Neubot has changed Internet address (anche, hence, ISP and/or location). To regenerate your unique identifier, you can run the neubot database regen_uuid command.

AUTHOR

Neubot authors are:


Simone Basso                  <bassosimone@gmail.com>


The following people have contributed patches to the project:


Alessio Palmero Aprosio       <alessio@apnetwork.it>
Antonio Servetti              <antonio.servetti@polito.it>
Roberto D\(aqAuria               <everlastingfire@autistici.org>
Marco Scopesi                 <marco.scopesi@gmail.com>


The following people have helped with internationalization:


Claudio Artusio               <claudioartusio@gmail.com>


COPYRIGHT

Neubot as a collection is:


Copyright (c) 2010-2013 Nexa Center for Internet & Society,
    Politecnico di Torino (DAUIN)

Neubot 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 3 of the License, or (at your option) any later version.

SEE ALSO

o  http://www.neubot.org/
o  http://github.com/neubot/neubot
o  http://twitter.com/neubot
Search for    or go to Top of page |  Section 1 |  Main Index


Neubot 0.4.16.9 NEUBOT (1) 2013-04-23

Powered by GSP Visit the GSP FreeBSD Man Page Interface.
Output converted with manServer 1.07.