Agents

Description

Soffid agents are the tool that allows the connection between the Soffid console and the target systems. To establish the connection with target systems, Soffid provides a large number of connectors that will be able to set up into the Soffid console.

You could see the complete list of Synchronization Server Connectors. 

Soffid administrator has the chance to easily customize attribute mappings for some connectors addons, without having to code it using Java. Soffid provides a graphical interface to perform attribute mapping.

An agent will appear disabled when this agent won't have a server assigned. Bear in mind to select the “Disabled” flag on Server URL criteria when you will query if you want to search for disabled, but defined agents.

Related objects

1. Synchronization server
2. Server
3. Account naming rules
4. User type
5. Password policies

Standard attributes

Basic

• Task engine mode: shows the current task engine configuration. For more information visit the Smart engine settings page.
• Name: agent's identifying name.
• Description: a brief description of the agent.
• Type: identifies the connector type to use. Different implementations of the server plugins included in the connectors installed into Soffid. Each type has a Java class bound, the name of the Java class implementing the connector is displayed next to the connector name.
• Server: synchronization will be performed with the selected server. It is allowed to select two servers in cases high disponibility will be necessary. If you choose two servers, when one fails, the other will be used.
  • If “Each main synchronization server” is selected, the agent will be run by every sync server.
  • If "-disabled-" is selected, the agent will be disabled. 
  • If you select a single sync-server, the agent only will be run in that server.
• Shared Thread: if it is enabled, the same thread will be shared to several synchronization servers. 
• Dedicated Thread: if "Shared thread" is disabled, it will be available the option to choose the number to threads to dedicate to the synchronization process.
• Task timeout (ms): add a timeout to the synchronization server tasks (query, insert, update, delete, update password, etc). If you add a timeout, when the connection gets this timeout, the synchronization server will stop the request and add it to the queue for a new retry later.
• Long task timeout (ms): add a timeout to the reconciliation server tasks (user, group, role, account, grants, etc). If you add a timeout, when the connection gets this timeout, the synchronization server will stop the request (no retry is added).
• Trust passwords: check it if you can trust on it to propagate their passwords to Soffid. Trusted password agents differ from the non-trusted in:

  • Temporary passwords generated from the console only propagate to agents that have trusted passwords checked. In the other case, the agents only receive definitive passwords.

  • When a password has reached its expiry date, it will automatically be disabled on agents where trusted password is not checked, so the user can no longer access it.

  • When the managed system detects a change in the user request password, the password will be propagated to Soffid only if the agent associated trusted password is checked.

• If you want to forward the authentication requests to trusted target systems, you must enable the Trust passwords option and the proper feature on the Authentication page.
• Authoritative identity source: check if the agent will be used as the source for users information. It is usually checked for the first load of users into Soffid, and then it is unchecked, being Soffid that manages users. Optionally, you can select a custom workflow to process incoming changes. 
• Read only: if it is checked, no change will be applied to the managed system. Only read operations will be allowed.
• Manual account creation: Check it if you don't want Soffid to create automatically new accounts for the user.
• Role-based: when "Manual account creation" is not checked (option selected is No), it will show "Role-based". Check it if only users with any role on this agent should be created. When the identity or account loses its permissions, the account will be disabled. Uncheck to allow users with no role on it.
• Groups: when "Manual account creation" is not checked (option selected is No), it will show "Groups". Identify the business units that are allowed to have an account on this system.
• User domain: it is the rule used to determine how to generate account names.  If the account name is the same as the user name (as it is normally the case), the “Default user domain” should be used. The user domain values are defined on the Account naming rules page.
• Password domain: determines the password policies that will be used. If the "Default password domain" is selected, Soffid passwords will be shared with the managed systems. The user domain values are defined on the Password policies page.
• User Type: when "Manual account creation" is not checked (option selected is No), it will show User Type. Only users of the selected types will be created. Any change made in this field involves all accounts to be recalculated. New ones will be added to the repository and managed systems. Some accounts will get disabled if the owner user does no longer belong to any authorized user type.

When uploading authoritative data for identities from a managed system, firstly, users will be created in Soffid as indicated in the attribute mapping, and secondly, accounts will be created for the managed systems only if the agent option "Manual account creation" is not checked and only for User Types indicate.

Connector parameters

The custom attributes depend on the used plugin. 

Here you will find all the information needed about the available Soffid connectors to integrate external managed systems.

1. AWS Connector
2. CSV Connector
   • Customizable CSV file (CSV Connector type)
3. Google Apps Connector
4. JSON REST Web Services Connector
5. LDAP Connector
6. Oracle Connector
7. Oracle EBS Connector
8. SAP Connector
9. SCIM Connector
10. Shell Connector
    • Invoker interface
11. SQL Connector
12. Windows Connector
    • HOWTO SSL access to Active Directory
    • Invoker interface for Active Directory
13. Zarafa Connector

Integration flows

Some connector addons have associated integration workflows. On the Integration flows tab you can view the integration flows related to the agent. You also can view in detail the workflows and test them. 

Attribute mapping

The attribute mapping tab only appears when the agent allows such customization. Soffid administrators have the chance to easily customize attribute mappings without having to code them using Java. The administrator users can select system objects and the Soffid objects related, manage their attributes, and make either inbound and outbound attribute mappings.

There is an action that creates all the default mapping depending on the agent connector type. That option creates automatically system objects with their attributes and properties. Once created the default mapping, those can be customized as required.

Properties

Some agents require to configure some custom attributes in this properties section.

You could see these attributes in the "Properties" section of every single page on the Synchronization Server Connectors chapter.

Attributes

Each object mapping defines an agent object name and one bound Soffid object type.

The left hand side attributes are managed system attributes, so they are agent dependent that is being configured. The right side attributes are Soffid attributes and must be selected from an existing list. 

It is allowed to use bean Shell expression in the source when the mapping is one-way.

System attributes

A configuration agent must define objects types that can be created on it. Each object mapping defines an agent object name and needs bound Soffid object type.

At this column, the system's attribute name will be displayed.

Directions

At the center column, an arrow will show the direction the information flows.

When the information flows from the system (left), to Soffid (right), the left column name can be replaced by a bean shell expression. This expression will be evaluated on the system object prior to uploading it to Soffid.

When the information flows from Soffid (right) to the managed system (left), the right column can contain a bean shell expression that will be evaluated prior to provisioning the user.

Here are some examples:

for (group: secondaryGroups) {
  if  (group.get("name").equals(primaryGroup)) {
    return group.get("description");
  }
}
return null;

Soffid attributes

You can consult the list of Soffid attributes:

• User Object
• Account Object
• Group Object
• Role Object
• Grant Object
• Maillist Object
• Membership Object

When evaluatting any expression, either the system or soffid attributes are available as script variables. More over, the following variables are available:

Triggers

It is allowed to define BeanShell scripts that will be triggered when data is loaded into the target system (outgoing triggers). 

The trigger result will be a boolean value, true to continue or false to stop.

A configuration agent can configure triggers related to the operation to be performed. There are different trigger type, that determines the specific moment on which the script will be triggered.

Triggers can be used to validate or perform a specific action just before performing an operation or just after performing an operation into target objects. 

Load triggers

On the Load trigger tab, it is allowed to set up a specific configuration for the agent and define BeanShell scripts that will be triggered when data is loaded into Soffid (incoming triggers).

• Full reconciliation: switch off to enable incremental load process and disable Soffid object removal.
• Propagate changes: switch off to prevent sync-server to create synchronization tasks after loading incoming changes.

To add a new trigger, it is mandatory first of all, to select a Soffid object on which the action will be performed. Then to select the trigger, that determines the moment on which the script will be triggered. Finally to define the BeanShell script that will be executed. The available objects are the following:

• User
• Account
• Group
• Role
• Grant

Triggers can be used to validate or perform a specific action just before performing an operation or just after performing an operation into Soffid objects. The trigger result will be a boolean value, true to continue or false to stop.

In a Load Trigger it is not possible access to mapping definitions configured on the attribute mapping tab. It will be necessary to use newObject{"attributeName"} to get the new value, or oldObject{"attributeName"} to get the old value

Massive actions

Provisioning all users on to managed systems

One of the main features in identity and access management (IAM) is automated user provisioning.  User provisioning is the process that ensures the users are created, with proper permissions, updated, disabled, or deleted on to managed systems.

All managed systems must have an agent configuration, which will determine the way to perform the provisioning.

Soffid shows information about the last time that the option was run and a report with the details. You can access to the report by clicking the verification icon (✓).

Propagate groups to agent

This option allows pushing to the managed system all the defined groups in Soffid. 

Soffid shows information about the last time that option was run and a report with the details. You can access to the report by clicking the verification icon (✓).

Reconcile (load target system objects)

The main purpose of reconciling process is to provide a mechanism to ensure that all users are aligned on the specific roles and responsibilities. Reconcile process discovers new, changed, deleted, or orphaned accounts to determine user access privileges.

Not every system connector has the capabilities needed to execute the reconcile process.

When "Read only" property, in Basic parameters, is checked (selected value is Yes), the reconcile process only considers unmanaged accounts. 

Soffid shows information about the last time that the option was run and a report with the details. You can access to the report by clicking the verification icon (✓).

Load authoritative data for identities and groups

Identities use to live on authoritative identity sources and they do in Soffid as well. Each identity may have any number of accounts on each managed system.

When "Authoritative identity source" is checked (option selected is Yes) Soffid will show the option that allows load authoritative data for identities and groups. 

That option performs the operations to load data of groups and data of identities from the managed system into Soffid,  following the rules configured in the agent.

Soffid shows information about the last time that the option was run and a report with the details. You can access to the report by clicking the verification icon (✓).

Also, Soffid create a parameter on the Soffid parameters page, with the information about the version of the data. If you need to perform the load authoritative action, it will be mandatory delete this parameter before perfom the action.

Generate target system potential impact

That option allows you to generate a report with all the potencial changes that would be perform on the manged system with the current agent configuration

If that option was perform previously, Soffid will show information about the last time that the option was run and the report with the potential impact. You can access to the report by clicking the verification icon (✓).

Account metadata

Accounts are default objects in Soffid. Agents allow you to create additional data, on the "Account metadata" tab, to customize the accounts created for that agent.

Standard attributes

• Code: short name used by scripts and connectors to access the underlying information. It is suggested to use short names without blanks or special characters to make it easier to use.
• Label: text displayed just beside the attribute value. It is advised to use short descriptions in order to keep the screen cleaner.
• Data type: The attributes can have different data types
• Prevent duplicated values: mark this field as a unique key for the object type. There is no chance of two objects with the same attribute value. Soffid smart engine will avoid the creation of duplicated objects.
• Multiple values: some attributes can contain multiple values for the same object. For instance, an attribute containing the languages a user can speak can be multi-valued, as a user can speak multiple languages.
• Maximum number of rows to display: when an attribute is multivalued, the screen size can grow a lot. To prevent such a big form, the system will only display a maximum number of values, and a scroll bar will appear to browse through the attribute values.
• Size: primarily for string attributes, specify the maximum length in characters of the attribute value.
• Values: primarily for string attributes, you can specify the allowed values for the attribute. Then, the text box that the user has to fill in the data will be replaced by a drop-down list.
• Visibility expression: write an optional BeanShell expression to check if the field should be displayed or not. The expression should return true or false. The following variables are exposed to the expression:

  • ownerObject: current object owning the attribute.

  • value: current attribute value.

  • requentContext: tip about the screen using the attribute.

  • inputField: the ZK input object (ZK Framework).

  • inputFields: a map to get access to any other ZK input object (ZK Framework).

  • serviceLocator: locator to use any Soffid engine microservice.

• Validation expression: write an optional BeanShell expression to check if the field value is acceptable or not. The expression should return true if the value is acceptable. If the expression returns false or any other object, a warning message will be displayed. When the expression returns a string value, the return value will be considered the warning message to present to the end-user. The following variables are exposed to the expression: 
  • ownerObject: current object owning the attribute
  • value: current value to evaluate.
  • requentContext: tip about the screen using the attribute
  • inputField: the ZK input object (ZK Framework).
  • inputFields: a map to get access to any other ZK input object (ZK Framework).
  • serviceLocator: locator to use any Soffid engine microservice.
• onLoad trigger:  write an optional BeanShell expression that will be executed just after preparing the user interface. The script can modify in any way the inputField object before it is displayed, but cannot modify other input fields.

  The following variables are exposed to the expression:

  • • ownerObject: current object owning the attribute
    • value: current value to evaluate.
    • requentContext: tip about the screen using the attribute
    • inputField: the ZK input object (ZK Framework).
    • inputFields: a map to get access to any other ZK input object (ZK Framework).
    • serviceLocator: locator to use any Soffid engine microservice.
• onChange trigger: write an optional BeanShell expression that will be executed just after the user has changed the object value. The script can modify in any way the inputField object or any other input fields.

  The following variables are exposed to the expression:

  • • • ownerObject: current object owning the attribute.
      • value: current value to evaluate.
      • requentContext: tip about the screen using the attribute.
      • inputField: the ZK input object (ZK Framework).
      • inputFields: a map to get access to any other ZK input object (ZK Framework).
      • serviceLocator: locator to use any Soffid engine microservice.

Actions

Agents query actions

Agent detail actions

Integration flows

Attribute mapping

Load triggers

Massive actions

Account metadata

Scripting

In the agents configuration it may be possible to use scripting to include logic in the attribute mappings and in the trigger scripts.

In the attribute mapping, if you use a script on one side, it will be mandatory a single direction to the other side:

• System attribute <= script
• script => Soffid attribute

Below an easy script to send a full name to the system:

system attribute <= return firstName + lastName;

Below a more complex script to create a main domain if it doesn't exist in Soffid:

String mailDomain = null;
if (email != void && email != null && email.contains("@")) {
    String[] mailTokens = email.split("@");
    mailDomain = mailTokens[1];
}
com.soffid.iam.service.MailListsService service = com.soffid.iam.ServiceLocator.instance().getMailListsService();
com.soffid.iam.api.MailDomain domain = service.findMailDomainByName(mailDomain);
if (domain==null) {
    domain = new com.soffid.iam.api.MailDomain();
    domain.setCode(mailDomain);
    domain.setDescription(mailDomain);
    domain.setObsolete(new Boolean(false));
    domain = service.create(domain);
}
return mailDomain;
 
=> mailDomain

ou could find a set of sample scripts: Sample scripts

You could find a link with the SCIM Query Language used in some methods as findUserByJsonQuery("query"). You can visit the SCIM chapter.

Below you could find a set of custom utility classes: Utility classes

More information

Password synchronization

The passwords a user has on an agent will be synchronized with any other "single user account" the user has on this agent. Shared accounts will never get its password synchronized.

Password in an agent will be also synchronized with any other account the user has on other agents that are sharing the same password domain.

The password change can be produced by an operator using the Soffid console, the user itself using the Soffid Self Service portal, or a timed automatic task. Furthermore, some managed systems can forward their password to Soffid in order to get them synchronized. In order to accept this password changes coming from managed systems, the trusted passwords box must be checked for the source agent.

Mind that this is the flow for normal user passwords. Temporary passwords generated by the Soffid console will only be sent to agents marked as trusted. Agents not checked as trusted will have a random new password instead. Later, when the user changes the password on Soffid or any trusted system, the new password will be notified to Soffid by the managed system, and every agent on the same password domain will actually get the new password.

Agents account management

The agent configuration sets the way accounts are created and disabled.

Whenever a user is modified, the following rules will be applied to check if the user should have or not an account on this agent:

1. The user type is checked against valid user types.
2. If there is a business unit or group bound to the agent, the user membership will be assessed.
3. If the role based box is checked, the system will verify if the user has any role or entitlement assigned on this agent.

If the user does not apply for any of the conditions, every account the user has at this agent will be changed to Disabled status.

If the user verifies every one of the conditions, the user can have an account on this agent. Every account the user has at this agent will be changed to Enabled status.

Unless the "Manual account creation" is checked, if the user can have an account on this agent, but it has no one, the account creation method will be invoked. To create it, Soffid will search for the user domain bound to this agent and will follow its configuration. If the user domain is configured with a script, this script will be executed and the result value will be accepted as the new account name. Mind that if the script returns a null value, no account can be created. 

If the returning value from the script clashes with an existing account, the existing account will remain unchanged, unless the existing account is marked as an unmanaged account. In such a case, the account will be changed from unmanaged state to single user.