Authentication and authorization
API docs
You are currently viewing the documentation for the unreleased version 5.0.0.CR2 of Vert.x. Visit the latest stable version of this page.

Mongo Auth Provider

We provide an implementation of AuthenticationProvider which uses the Vert.x MongoClient to perform authentication and authorisation against a MongoDb.

To use this project, add the following dependency to the dependencies section of your build descriptor:

  • Maven (in your pom.xml):

<dependency>
  <groupId>io.vertx</groupId>
  <artifactId>vertx-auth-mongo</artifactId>
  <version>5.0.0.CR2</version>
</dependency>

  • Gradle (in your build.gradle file):

compile 'io.vertx:vertx-auth-mongo:5.0.0.CR2'

To create an instance you first need an instance of MongoClient. To learn how to create one of those please consult the documentation for the MongoClient.

Once you’ve got one of those you can create a MongoAuthentication instance as follows:

MongoClient client = MongoClient.createShared(vertx, mongoClientConfig);
MongoAuthenticationOptions options = new MongoAuthenticationOptions();
MongoAuthentication authenticationProvider =
  MongoAuthentication.create(client, options);

Once you’ve got your instance you can authenticate and authorise with it just like any AuthenticationProvider.

The out of the box config assumes the usage of the collection with name "user", the username stored and read in the field "username", and the password stored in the field "password".

In order to avoid duplicates of user names your "user" collection should have a unique index on "username". In order to do this you should run the following snippet on your mongo server:

db.user.createIndex( { username: 1 }, { unique: true } )

The reason you should add the index is that due to the nature of mongo doing a query first to verify if a username is already taken and then insert a document cannot be run as an atomic action. Using the index the code will try to insert the row and fail if duplicate.

You can also change all the defaults for the mongo collection and column names using any of the methods:

setCollectionName setUsernameField setPasswordField if you want to adapt that to your needs.

The default implementation assumes that the password is stored in the database according to the generic HashingStrategy (so the salt is stored in the hash).

Vertx Auth JDBC and GDPR

GDPR is a regulation from the common European Union law. It overrides/supercedes national data protection laws and extents the previously existing directives. This section of the manual is by no means a thorough walkthrough of the regulation, it is just a small summary how this component adheres to the requirements. Companies not adhering to the equirements can be fined on 4% of the turnover or 20 million euro. Therefore we want to make sure that as a user of Vert.x Auth JDBC you’re are on the good track to comply.

The law defines certain terminology:

  • Data Subject - Person whose personal data is processed (e.g.: User)

  • Personal Data - Any data about an identifiable or identified person

  • Data Processing - Any operation (manual or automated) on personal data

  • Controller - The entity (company) that requests and uses the data

  • Processors - Any entity that processes data on behalf of a controller (e.g.: cloud service provider)

GDPR defines the following functionality:

  • "Forget me" - Right to erasure

  • Mark profile as restricted - Right to restriction of processing

  • Export data - Right to portability

  • Allow profile editing - Right to rectification

  • See all my data - Right to access

  • Consent checkboxes

  • Age checks

  • Data destruction - Data minimization principle

This module complies to the GDPR law by not storing any identifiable information about a data subject. The only reference is the username which is not linked to any personal data.

In order to add personal data to your application you should create your own data schema and use the username column as a reference to your data. As a tip you should have a boolean flag to mark the personal data as restricted to comply to the right to restriction of processing which means that if you need to handle the data, e.g.: send a bulk email from a mailing list you are not allowed to do so if the flag is true.

The right to erasure does not mean that you must wipe all records from your application, e.g.: in a bank this right cannot be used to erase a running loan or debt. You are allowed to keep your application data but must erase the personal data. In case of Vert.x Auth JDBC you should delete your table but can still use a reference to the username as long as is not possible to link the username to the personal data.

Important note is that this must survive backups! As a tip backup the data, and data erasure on different archives so they can be replayed individually.

Authentication

When authenticating using this implementation, it assumes username and password fields are present in the authentication info:

Credentials credentials =
  new UsernamePasswordCredentials("tim", "sausages");

authProvider.authenticate(credentials)
  .onSuccess(user -> System.out.println("User: " + user.principal()))
  .onFailure(err -> {
    // Failed!
  });

Instead of the username and password field names used in the previous snippet, you should use: getUsernameCredentialField and getPasswordCredentialField

Authorisation - Permission-Role Model

Although Vert.x auth itself does not mandate any specific model of permissions (they are just opaque strings), this implementation assumes a familiar user/role/permission model, where a user can have zero or more roles and a role can have zero or more permissions.

mongoAuthZ.getAuthorizations(user)
  .onSuccess(v -> {
    if (PermissionBasedAuthorization.create("commit_code").match(user)) {
      // Has permission!
    }
  });

If validating that a user has a particular role then you simply use the RoleBasedAuthorization.

mongoAuthZ.getAuthorizations(user)
  .onSuccess(v -> {
    if (RoleBasedAuthorization.create("manager").match(user)) {
      // Has role!
    }
  });