SAML 2.0 best practices for custom user fields - single-sign-on

We have an existing web application which works with its own users, stored in a local SQL server database and each user has specific rights to access specific content.
A new customer is interested in using this application, but their requirement was that it should have SAML 2.0 integrated.
Now, I'm new to SSO, but I assume they want this so all user management is done at their place and that their users can log on at any of their services, using the same credentials.
The user rights defined at our web application however, are application specific; they have no meaning outside our application. So I assume these fields should remain stored and managed at our end, instead of managed at the identity provider?
But what would be the best way initiate this? I could create an empty user record locally whenever the user logons for first time, but he would literally not be able to do anything without proper rights. I could give him default right settings, but there should be at least 1 administrator to distribute the rights of the remaining users. Is it common to ask additional fields at the identity provider to initiate users for a specific service?

This can be done using entitlements sent in a SAML2 attribute. e.g. if you have a normal user profile for your application, you can ask the IdP to release an attribute such as eduPersonEntitlement or urn:oid:1.3.6.1.4.1.5923.1.1.1.7 as it's known in SAML2. If the IdP owner agrees to release this attribute you can agree a controlled vocabulary with them for the value of the attribute.
So when a normal user logs in to your application using SAML2 and their attributes arrive in an AttributeStatement with a collection of SAML2 Attribute instances you can look for urn:oid:1.3.6.1.4.1.5923.1.1.1.7. If it has a value of, perhaps, https://yourapp.com/entitlement/user you can set them up with a normal user account. If they have a value of, perhaps, https://yourapp.com/entitlement/admin, you can set them up with admin rights.
Just be careful to compare what they already have (if anything) with their incoming SAML2 entitlement, in case they longer are an admin. The IdP decides who is a normal user and who is an admin but that's up to them to work that out. All you care about is the value of urn:oid:1.3.6.1.4.1.5923.1.1.1.7
The full set of eduPerson attributes is here

Related

Organize and validate companies in keycloak

I want to use Keycloak for my application. Since our current data model expects a user name, password and the abbreviation of a company, I would like to know if there is a way to add an additional parameter to the login screen in Keycloak where the abbreviation of the company has to be specified. During the authentication process it should be checked whether the user belongs to this company and if so, the token which is created should contain the company's abbreviation.
So far I manged to modify the login theme like this:
Login screen
I know it is possible to add additional information to a token via client scopes and attribute mappers, but as far as I know, this is only possible for single users.
Also I'd like to know, which is the best way to administrate companies in keycloak.
Thanks in advance.

How should mapping of service provider users to identity provider users work?

I'm working on setting up a service provider that supports SAML. I've added two identity providers - one custom one that I built from SimpleSAMLphp and then ssocircle. So I log in to the selected identity provider, and then it returns to my service provider and I inspect the attributes of the SAML Auth object. The identity provider I built returns whatever I want it to (obviously). The ssocircle one only returns e-mail, first, and last names.
So now to map this to the user of the service provider, I have to use some value the identity provider provides. So this led me to wonder how it should be done. Since ssocircle only gives me e-mail as a useful value, do I just use the e-mail to map to the SP user?
Let's pretend for a second that ssocircle doesn't validate e-mail addresses. So now if I create a second account at ssocircle with the same e-mail, I can log in as my coworker who I know has admin privileges.
So my question is, do I handle this? Or is the onus on the admin who set up the identity provider and say "well you shouldn't have used an identity provider that doesn't validate e-mail addresses!" or something of that nature? Or should I only allow identity providers that pass a certain value, like userid or 0.9.2342.19200300.100.1.1? Is there something that identity providers commonly use?
Well, you said it, two different identity providers. They both should be passing not only the email but different entity ids and certificates.
In multi-tenant applications that would mean two different applications, but if you plan to allow multiple IDPs to point to a single app you will need to ensure that same email but different entityID create two different users and or throw an error after the first was created that the second cannot be provision nor access.
Interesting question. These days people think always of auto federation of users by some attribute. In early SAML federation days federating two unrelated users was a manual step in which a user logs in at the IDP and logs in to the SP providing both sets of credentials and then manually federated these two user accounts. The process guarantees that only the user who has access to the accounts at the IDP and SP controls the linkage between the two. It also allows anonymous naming identifiers (SAML persistent NameIDFormat) which protects privacy because even the IDP does not know the user name at the SP and vice versa.
Unfortunately the process was to complicated for users and with the success of OpenID the aspect was getting less and less important.
To answer your question: What you describe happens in the real world -see Office 365 authentication bypass
You need to check that the IDP is authoritative to send a specific attribute and attribute scope in case of two IDPs.
In case of one IDP the attribute must be verified (SSOCircle verifies email address) and it should better be unique (For example SSOCircle userId) to avoid that two users with the same attribute are mapped to a single user at the SP.
If the userid's are not the same (e.g. you use a simple user ID at the IDP and email address format at the SP) you can still add a correlation attribute at the SP (e.g. an attribute named ssocircle-userid) and use that to link the user accounts.

SSO - Do we always need to keep user database at IdP and SPs in case of role based application

I'm planning to use simplesamlphp as an IdP. I've reasonable amount of user database who "can" share 5 different applications.
Currently, user have to go on each application and put username and password which is surely not what I desired. The solution is to keep separate authentication server and use SSO to serve 5 different applications (as service provider).
Although, the application can share same user database but the services are completely different and they heavily rely on user roles. Let's take an example of two applications: SP1 and SP2
SP1 is in and laravel and heavily role based system. At any point of time, one user can have 5-6 roles depending on which the application functions / provides information.
SP2 is completely different it could be just a blog site in drupal where we want to collect the user data that's it. We will have a minimal role requirements here.
When user comes on SP1 and tries to get restricted page then, they will redirect to IdP which can authenticates user. If user does not exist in IdP, it will return false; otherwise returns back to SP1 with some relevant attributes. Let's assume, for SP1 this is a newest user, we've an option to save that user in SP1 database (in fact we should bind that user with SP1 specific role) but then, we're storing user in two places: IdP and SP1.
...And the same goes for SP2. In this case we are storing user in total 3 places: IdP, SP1 and SP2. The only benefit we got is, user didn't have to enter username and password to SP2 it automatically handled by our IdP.
Questions
Do I need to store user in both SP1 and SP2? If not, then how can I handle all role specific functionalities.
If not, then how do I initialize the user session and assign SP specific roles? based on SAML attributes?
SAMLv2 has not concept of authorization, it's a technology for cookieless web-based SSO. XACML would be a standards-based technology for authorization / fine grained access control / entitlement.
Most application are lacking the ability to map the internal RBAC system to something external.
If you rely on information within the SAML assertion, like attribute statements (typically the way this is done), then you need to have some way to map the attribute values to the roles of your app. If the app can not be changed to implement such mapping you need to provision identities and role assignments in some way.
Typically the IdP implementations allow to specify different attribute mappings per SAML SP entity.

SAML: group memberships

I was told that it is possible to give information about group membership during a SAML authentication request. We have to connect to an application that does use SAML (we are at the end that is creating the SAML answer XML). Authenticating a user works fine but I can't find a way to specify a "member of" or similar attribute.
Can you explain to me how to pass group memberships in SAML during authentication or have an example ?
I know there is a possibility to take care of authorization in SAML at a socalled Policy Decision Point. But this would mean that a SAML flow would happen for each or some (if batched) entities we want to check authorization for.
Let me give you an example what we are trying to do. This example is made up but shows the problem we want to solve.
Let's assume you have a hard drive with lots of directories and files on it. We use SAML to authenticate the person that tries to access that drive. Members of the group "admin" are allowed to read and write and members of group "user" are only given read permission.
Because of this we want to send the group memberships of a user when he or she authenticates. Because otherwise it would mean that the application has to check for every file if the user is in the necessary group. If it was clear from the beginning (after authentication took place) that someone is a member of a specific group the application can cache that in memory.
As with most things in SAML, this depends on the identity provider as well as the application.
Many identity providers have access control to allow or deny specific users or groups to access an application. Sign-on is then denied by the IdP after authentication based on the ID of the application (from the SAML request) and access control configured in the IdP. If you just want to control who has access to the application, then this is all you need, and most, if not all, identity providers should have you covered.
Sometimes you also want different users or groups to have different permissions inside the application, or you want memberships in certain groups on the IdP to be mapped to membership in groups inside the app's own user and group database (or other application-specific group-like concepts). There is no standard for this, but some identity providers allow the definition of attribute mapping rules based on group membership. For example, in ADFS, you can create a claim rule that sets a specific SAML attribute to a specific value if and only if the user is a member of a specific group in AD. You would then set up such a rule and set a SAML attribute that your application understands to a value it understands.
Alternatively, if an application supports configuration of multiple IdPs, then you can configure one app / IdP mapping per group, and for each group add an attribute mapping rule with a hard-coded group name in an attribute that your application understands.

How to setup initial user/role enabling SAML SSO

We are a service provider. Suppose in our application, we originally have our own user/role management. Different users with different roles are allowed to use different features. So that when a user login we need to know which roles this user has, and prepare appropriate UI. We have administrator role, users with this role can assign roles to other users.
We are thinking of enabling SAML SSO for our application, now the problem is how do we setup roles for each user.
Solution 1, we relies on IdP to provide role information for each login user, the role information may come along with Assertion, but this may not work for all the IdPs.
Solution 2, we only retrieve user from the IdP, and manage the roles in our own application. For example, when we get an Assertion, we retrieve the username(or email address), and match with a record in our DB, if it doesn't exist we automatically create one for this new user. Then we rely on users with administrator role to assign correct role for this new user.
Now the questions is where is the first administrator coming from? Our customer gets our application, and turns on SAML SSO, now there is no users in the DB yet, then how can we resolve such bootstrapping issue? Is there any kinds of standard way? We have come up with different options but not sure which is better and what are the concerns for each options.
Option 1, have a default built-in administrator user. There is a regular native login page that built-in users can login without going through IdP(there is an option to turn it on/off if SAML SSO is enabled)
Option 2, during SAML SSO setup, ask for the administrator user name, so that we automatically create this user in our DB with administrator role. Then when this user login through IdP we could match him in our DB.
What are the other options?
For your first question about who should handle the roles. As I understand every customer has your One of your service provider software. And it connects to a central IDP that you own. If this is the case, it feels like its to complex letting the administrators handle roles on your IDP. I would go for number two.
About question number two. I have been in the same situation I can not remember that I have seen a obvious standard solution for this.
What we did was option 2. It works fine but it adds some complexity to the install procedure. We choose it because we would not have a native login page.
I thing maybe I would be better to go with option 1.
You can get the roles from the different directories of your IDP using your authorization layer. It removes the constraints on SAML and gives the same results as your solution 1.