I got absolutely confused when trying to understand the differences between MicrosoftAccount, AzureAD and OpenIDConnect authentications.
I am trying to build a .Net Core MVC app that allows some users to authenticate with local accounts, but some with Microsoft accounts. I do need to have a local user in DB for both types of authentications as I have some custom authorization mechanisms built on that.
I started with creating the app from template and selected "local accounts" authentication. Then I added the MicrosoftAccount authentication according to this tutorial (https://learn.microsoft.com/en-us/aspnet/core/security/authentication/social/microsoft-logins?view=aspnetcore-3.1). This is using Microsoft.AspNetCore.Authentication.MicrosoftAccount and seems to be working fine.
However, when I create a new app from template and select "work or school account" authentication I can see it uses a different library - Microsoft.AspNetCore.Authentication.AzureAD.UI. It seems to do the same thing. I can see there are events I could hook into to connect the AAD user with my local DB.
Looking through the web I found some other tutorials that were using OpenIDConnect for the same purpose.
How are those methods different? Which one should I use and why?
In future I would like to be able to query the user's directory for a list of other users. Would that requirement be easier met with either of those three methods?
MicrosoftAccount: This is the login with a general Microsoft account, using OAuth2. This is also what Microsoft will refer to as “private account” and useful when you want to authenticate someone just using their Microsoft login as an external authentication provider. This is similar to how you would sign in to sites with your Google account.
AzureAD: This is the sign in to a specific Azure tenant. This is often understood as a “work or school account” because it doesn’t sign people in using a public account but rather some account bound to some organization. Usually, you would have an Azure tenant where you have direct users (or configure it to allow external users) but you want to control access through that Azure AD.
OpenIdConnect: This is the general OpenID Connect protocol which you can use to sign in with many different authentication providers because it is a protocol that many of them will support. You can use OIDC to sign in to either of the above (and many other services) but that will require you to do some more configuration as you will need to figure out specific addresses for example.
You can always use the OpenIdConnect or the OAuth authentication scheme to authenticate with most authentication providers but those are the “manual” schemes which will require you to configure additional things. All the other authentication schemes, including MicrosoftAccount and AzureAD but also the other ones like Google or Twitter build on top of those protocols and come preconfigured so that you do not need to set up much else. So those are mostly for convenience and for more specialized support.
So when you want to authentication through Microsoft or Azure, then you should choose MicrosoftAccount or AzureAD. Which of those depends on where you want to authenticate with. If you have an Azure AD, then you should use that.
Related
I have a need to implement Open ID Connect in an ASP.NET Core 5 web app for authentication at an organization. I cannot test against the identity provider as this org has their own internal one.
I thought Open ID Connect (oidc) was a standard - yet when I look at docs and sample code for the various providers around, they all either have something provided with ASP.NET or I have to install their package.
For example, to use Google, the ASP.NET Core docs say to use services.AddAuthentication().AddGoogle(). For Facebook, it says to use .AddFacebook().
For Auth0, it wants me to install the package Auth0.AspNetCore.Authentication.
Is it not possible to just add OIDC authentication to my app and have it work with any OIDC provider and just change a configuration file to specify the authority URL, Client ID, Client Secret?
I'm confused about why I need these provider-specific calls and packages.
Architecturally, tieing an app to a single form of authentication is entirely wrong, as you suggest. These packages have limited use, perhaps for very simple use cases.
The original OAuth 2.0 spec from RFC6749, from 2012, describes how the OAuth framework is designed, to externalize difficult security from your apps:
The client only implements a code flow
It redirects to an authorization server (AS)
The AS can authenticate users in a myriad of potential ways, including many identity providers
After authentication (and possibly consent) the AS issues tokens to your apps
Tokens enable authorization in your APIs in a myriad of potential ways
Once you've implemented a code flow in your app, your set of users can potentially login in many ways, with zero code changes in the app:
Password sign in (a default option)
Multi-factor authentication (in a dynamic way potentially)
Cloud platform identity providers (for engineering staff)
CRM identity provider (for client focused staff)
SAML identity providet (for users from business partners)
Webauthn, Passkeys and Digital wallets (for some of your customers)
Unless you have a very good reason, stick to OpenID Connect standards based code flows in clients. You can do all of the above using the Microsoft libraries. Auth0 have good libraries also - just make sure you use the standards based ones.
OpenID Connect is an open standard that organisations use to
authenticate users. IdPs use this so that users can sign in to the IdP
From this blog.
And about the OIDC protocol, it allows you to authenticate users, and is designed for users to sign in many websites with only one account, usually it's a social/work account. This is only a protocol and you have to use an implement such as Google/Azure authentication to allow your users to sign in with their specific account.
By the way, since the implements are from different companies, so the configuration in our codes are different and they required different nuget packages like Microsoft.Identity.Web. For example, when using Azure, we need to set such as client id, client secret, tenant id, domain, redirect url... in appsettings.json.
I'm trying to setup an application to validate identity using Azure AD and acquire a token to allow access to a secure api. The front end application is written in angular and allows anonymous access. What can I use to access AAD authenticate and return an access token?
This will be an angular 6+ UI that is communicating to a secure .Net api using Azure AD for authentication. I have done a couple days research and everything points to a user logging in to authenticate using the login page. I need it to be by app and open the login page. I tried a couple examples where it utilized authentication/authorization and that didn't work because the app needs to authorization the user to talk to the api. I have seen where people were using Microsoft graph but once again it was user based and they were redirected to an azure login. I am looking for a solution that will allow me to setup an account in azure ad and authenticate the app on start to get an access token to allow communication to my secure api. If I have missed something somewhere in my research and test attempts let me know. This is my first Azure AD auth attempt and I feel like I am missing something for application authorization.
The problem is an Angular app is what we call a public client.
It cannot hold secrets and thus cannot prove its identity.
So, only user-based authentication flows should be used by public clients.
Confidential clients on the other hand can hold secrets as they run on servers that you control.
So for example, a back-end Web application or API would be a confidential client.
Those can use the client credentials flow to acquire access tokens and call APIs as themselves without a user being involved.
There is a bit of a fundamental issue in your question.
Your API requires authentication, but you want functionality to be available to anonymous users.
So you want to bypass authentication.
If you really want to bypass authentication for parts of the API, you could just make those endpoints available anonymously without a token.
We have a client with following existing product line
Website1 (Uses form authentication using user information stored in SqlDB for website1)
WebSite2 (Uses form authentication using user information stored in SqlDB for website2)
WinApplication1 (uses Active Directory to authenticate user)
More products in line for future release.
They want to implement Single Sign on (SSO) for all their products with following features-
Product list that will be accessing by a user will be depended on his/her role.
It should be configurable. Roles are being defined in DB having Fk relation
with user table.
SSO service should be flexible to accommodate new product.
should use SAML to sucure token.
WinApplication1 has link for website1 and website2. If an user is login in WinApplication1
and click on any links of website1/website2, user should able to access the sites without
entering his credential.
If user is logned in in any product that user should be all to access any other product
he has access right without enterting userid/password.
We have following query
We are planning to implement SSO using WCF and WIF. Can all above requirement be fulfill
by using WIF.
We want to keep user authentication part in respective product as each product has their
own user store. OR WE SHOULD HAVE CENTRALIZED USER STORE?
What shouldl be the responsibility of STS (implemented in WCF) in respect to SSO &
security?
We have read lot about WIF from MSDN(http://msdn.microsoft.com/en-us/library/hh377151(VS.110).aspx, http://msdn.microsoft.com/en-us/security/aa570351.aspx) could not find any solid article/sample giving some practical scenarios. Most of them about using local STS…may be it is my bad luck.
I found a article that matches most of your requirements with practical implementation example.
The article describes an easy approach for programming against Windows Identity Foundation in a Single Sign-On (SS0) scenario from the angle of a relying party application. WIF supports a variety of Claims-based authentication scenarios but this document will focus upon using WIF to develop an application that supports SSO given the use of a SAML 2.0 token containing some basic claims.
Article also includes some part explaining about using roles in AD Users:
http://www.c-sharpcorner.com/UploadFile/scottlysle/windows-identity-foundation-and-single-sign-on-sso/
To impersonate the Original Caller in WCF Calling from a Web Application, you additionally need to follow:
http://msdn.microsoft.com/en-us/library/ff650591.aspx
Update:
difference between IP-STS and FP-STS and how to choose where to put STSes in your architecture. You will learn about federation, home realm discovery and how to leverage the WIF extensibility model in order to handle multiple identity providers.
http://bit.ly/bTo3w4
How a generic web site can be enhanced with identity provider capabilities regardless of the authentication technology it uses, simply by adding an STS page.
Another lab shows how to use an existing membership store for authenticating calls to a custom STS and sourcing claim values.
http://bit.ly/bQAQhA
WIF and WCF
http://bit.ly/8X63T5
Hope it helps.
I am going to preface this by saying I am a novice in this field. Our system has a large number of users, some of whom, log in using SAML authentication, others use the forms authentication on the home page, and some use a custom built Single Sign On solution. This happens because different clients prefer different modes of authentication.
We are looking to streamline authentication so that it is not a function of the web application, and we are therefore going by the claims model. One of the infrastructure options to build the STS is the Thinktecture Identity server.
My question is this:
- Is is possible to have multiple STS for the same web app, and if so, where does the switching happen? I.e if user1 uses SAML, how is this user's credentials sent to that specific STS?
Also, is Thinktecture the right infrastructure to be building something like this, or should I look at something else?
I've been scratching my head over this issue for over a week. We have a web app that we would like to implement SSO for. SSO with windows active directories of our clients (i.e. we essentially need to authenticate against our clients' active directories without much trouble)
The only thing I am 100% sure about is that I will needed a security token service that will have to communicate with an Identity Provider. My question:
Which service is most suitable for the above scenario (AD FS? OpenID & OAuth 2.0? SAML 2.0 and shibboleth?)
How will I connect to the active directories of the clients? Maybe I'm not understanding how the STS is to be used, could anyone clarify? I'm working with an Azure Web App
Will there have to be a different IdP for each client? Will the client have to do more than just give us standard information? What would this info be?
...should I be using Windows Identity Foundation?
HELP :( ... this is an SOS
If anyone could clarify at all, I will forever be grateful. I normally upvote anything I find helpful and accept whichever answer is the best so feel free to answer with what you think might be useful in helping me understand how I can achieve what I am after.
These are the three options I know:
As you mention one option is ADFS this solution means that your customers should install and expose Adfs. ADFS means Active directory-Federation Services, so in this case your application needs to speak WS-Fed (not oauth). Typically if the user is inside the LAN adfs uses integrated auth, if not it will prompt credentials.
WAAD is a new service from Azure, it allows companies to expose their directories to use in cloud applications. With this approach your customers need an account in Azure, create a directory and use the dir sync agent. Your application will talk SAMLP with WAAD.
Auth0 is an authentication broker that allows developers to use social but also enterprise identity providers like AD but also google apps, waad, adfs, salesforce, etc. if your customer only has AD you will provide him an msi for a windows service, that will bridge the company AD with your auth0 account, you can have as many AD as you want. Your application speak oauth with Auth0. This agent supports kerberos authentication as well. The following graph explains this solution:
Disclaimer: I work for Auth0.
WIF doesn't support SAML or OAuth.
Your application is in Azure.
Suggest add WIF to the application and then "bind" to Azure Active Directory. In VS 2013, use the "Change Authentication" feature for this.
Make the application multi-tenanted.
Each customer has their own tenant. User DirSync to sync. each customer AD with their AAD tenant. (That gives same sign-on). Adding ADFS to each customer gives single sign-on.
However, the customers will probably push back on this because of perceptions around security.