I'm aware of how SAML is used for single sign on (SSO). That is, redirection to IDP from SP and getting the user's identity from the SAML response/assertion.
My question is - Does SAML 2.0 specification define how to pass username and password as part of a SAML request xml for authentication? Note that I'm not talking about single sign on and just want authentication of username/password.
Thanks,
The SAML standard supports passing a user identifier in the <saml:Subject> field of the <AuthnRequest> (i.e. the request for authentication).
There is however no built in support for passing a password as part of the AuthnRequest. IMHO doing so goes against the principles of SAML2 as that expects the Idp to only use a password when authenticating. Normally the Idp may use any means it finds suitable to confirm the identity of the subject. That could be a password, but also a certificate or a one time pad exchange over SMS. Or something else - it's up to the Idp.
That said, there is an <Extensions> element in the <AuthnRequest> that could be used to carry a password. Doing so would require careful security considerations as the AuthnRequest contents are not designed to be kept secret. If using the Http Redirect binding the contents are logged in a web server and visible in browser history. If using the Http POST binding the password is still visible to the browser. I would suggest using the SOAP or Artifact binding to make sure the data is transferred directly from the SP to the Idp. Note however that those bindings have considerable less support in frameworks.
Related
Trying to implement OpenId Connect in Web Application consisting of following components
Identity Provider
Resource server
Single Page Application acting as Client.
Identity Provider and Resource Server are the same application.
SPA use Password Flow to get access_token and stores into the cookie. Storing access_token into cookie has it's security threads, but's it's a different story.
Problem
access_token issued by IdP is expired after 30 min and SPA needs to renew token without asking users for credentials again.
Solution
IdP returns refresh_token along with access_token. Whenever SPA gets 401 from Resource Server, it sends refresh_token to IdP and get's new access_token back.
Problem
Sending refresh_token to SPA is bad practice.
A Single Page Application (normally implementing Implicit Grant) should not under any circumstances get a Refresh Token. The reason for that is the sensitivity of this piece of information. You can think of it as user credentials since a Refresh Token allows a user to remain authenticated essentially forever. Therefore you cannot have this information in a browser, it must be stored securely.
Suggested solution
When the Access Token has expired, silent authentication can be used to retrieve a new one without user interaction, assuming the user's SSO session has not expired.
I think Silent Authentication is not applicable to Password Flow when IdP and Resource Server is same application. access_token issued by IdP is only piece of information which can be used to authorize against Resource Server/IdP after its expiration, how a client can convince IdP to issue new access_token? (without sending refresh_token)
Found angular-oauth2-oidc library which uses refresh_token to renew access_token.
What is best practice/solution in this case to renew access_token?
technical details
Identity Provider - ASP.NET Core + Openiddict library.
SPA - AngularJs application.
Single page applications must not receive refresh tokens. That has been established rules in OAuth 2.0 and OpenID Connect.
One good option I see here is to use Implicit Flow. This will establish a front channel session from your browser to Identity Provider. With password grant type you do a back-channel call (POST), so you don't get such session.
Usually this is a cookie which points to information about previous logged in status (these are identity provider specifics). With completion of the flow, SPA will receive the access token. As you figured out, it will expire. But once that happens, SPA can trigger another implicit flow, but this time with prompt query parameter.
prompt
Space delimited, case sensitive list of ASCII string values that
specifies whether the Authorization Server prompts the End-User for
reauthentication and consent. The defined values are: none , login, consent and select_account
If you identity provider maintain a long lived session (ex:- few hours or days) or if it maintain a remember me cookie, SPA could use prompt=none making it to skip login step from identity provider. Basically, you are getting browser based SSO behaviour with this.
Using the Resource Owner Password Credentials flow defeats the refresh token storage argument: instead of not being able to store the refresh token in a secure place, the SPA would now have to store the Resource Owner credentials in a secure place (assuming you want to avoid requesting username/password from the user frequently). The Implicit grant was designed for usage with an SPA, so it is better to stick with that.
Further to previous answers, the latest OAuth working group guidance for SPAs no longer recommends use of the implicit flow.
If you have simple, shared domain app (IdP, RS and client on a single domain) then you should consider not using OAuth at all. From the doc:
OAuth and OpenID Connect provide very little benefit in this
deployment scenario, so it is recommended to reconsider whether you
need OAuth or OpenID Connect at all in this case. Session
authentication has the benefit of having fewer moving parts and fewer
attack vectors. OAuth and OpenID Connect were created primarily for
third-party or federated access to APIs, so may not be the best
solution in a same-domain scenario.
If you are using OIDC/OAuth in a SPA, they recommend the auth code flow with PKCE.
Alright, I've spent several days looking for a proper solution on how to properly authenticate users when working with SPAs.
I have my own website.
I have my own API.
I have my own Single Page Application.
I have my own database of users.
The Goal: I need to get an access_token by providing a username and a password.
I looked at OAuth2 Implicit Grant, but it requires users to Approve/Decline the app after successful authentication. It doesn't work in my case since I own both the app and the API.
I looked at OAuth2 Password Grant, which is not perfect since I need to expose client_id/client_secret.
The reason I'm looking at OAuth2 is because the API will eventually be public.
Is there a standard way of doing this? My current options:
Forget about OAuth2 and manually generate access_token when user POSTs username/password (in this case I'd have to introduce OAuth2 when API goes public)
Use OAuth2 Password Grant and inject client_id/client_secret on the server, so just to keep client app very simple (also avoid all of those dev/staging/prod client_id/client_secret pairs)
Implicit Grant
You are right that Implicit grant type does not look appropriate. But I think your reason for not favoring it is incorrect because the approval step is not mandatory and in Spring OAuth 2 implementation (I don't know which implementation you are using) you can configure the Authorization server to auto approve authorization requests so that the approval step is skipped.
The reasons I think the "Implicit flow" is not suitable are
The client authentication step by providing client secret and authorization code is missing. So less security.
The access token is sent back as a URL fragment (so that the token doesn't go to the server) which will continue to stay in browser history
If XSS attack occurs, the malicious script can very well send the token to the remote server
Resource Owner Password Credentials Grant
If the authorization server and the resource server are the same, I think this is a quick way of getting up and running. RFC 6749 in Section 4.3.2 says:
If the client type is confidential or the client was issued client credentials (or assigned other authentication requirements), the client MUST authenticate with the authorization server as described in Section 3.2.1.
This means that the client authentication with client secret is not mandatory here. Now, for authorization code grant type, we need the client secret because the user provides his/her credentials directly to the authorization server and then when the client requests for the access token, it doesn;t have anything else other than the client secret to prove to the authorization server that this is a genuine request.
But in case of resource owner password credential grant type, the user has provided its credentials to the client itself and the client will then send these same user credentials for requesting access token. Therefore, the access-token request can be authenticated with the user credentials only and if we don't provide a client secret here, I don't think we are losing anything in terms of security.
So, you can definitely use password credential grant type in your SPA.
Authorization Code Grant
I think this should be the preferred option provided the client secret is not stored in the browser. After user authentication (and optionally user approval), the authorization server can redirect the browser to a server side endpoint with the authorization code in the URL. The server side end point will the request for the access token using the authorization code, client id and client secret (which is stored in the server side only). Once the access token is available, the server side endpoint can redirect (HTTP response code 302) the user to the SPA URL with appropriate cookies for CSRF protection and access token. Thus we are not storing the client secret in the browser.
By using authorization code grant type, you are basically making the solution more secured and generic. In future, if you want to do a single sign-on with a different SPA, you can do that easily by reusing the same authorization server with its integration with the authentication database (preferably an LDAP server).
For further details, refer to my StackOverflow answer here.
Building off what has been said already, I would recommend the 'Authorization Code Grant' but with the addition of the PKCE (Proof Key for Code Exchange / 'pixie') extension - for added security, regardless of whether you're implementing a 'public' or 'confidential' type client.
With PKCE, you don't need a client-secret for public clients (/it's kind of like generating a temporary client-secret at the very outset/beginning of each authentication attempt/instance - although even with PKCE for confidential clients you should ideally still use a client secret).
Let's say I have a system that's currently like this:
Monolithic Web App: Contains its own accounts and relies upon clients to log-in with (essentially) HTTP BasicAuth. That is, the username & password are getting passed to the server.
Thick Client: Logs-in to the above app, receives access token it uses thereafter for REST API calls.
Basically, I'd like to transform the above into this sort of system:
SAML 2.0 IdP: System of record for identities
Same Web App, minus Authentication responsibility
Thick Client: Unchanged. <-- hard requirement
So, crucially, I can't have the thick client do the standard SAML 2.0 Browser SSO redirects. Is there any solution? Essentially, I'd like the same functionality as OAuth2's password_grant, but in the SAML 2.0 world.
Doing some research, I've come across the SAML Enhanced Client or Proxy, but support seems spotty. The frustrating thing is, I have the darn credential in cleartext at the WebApp; is there some simple way to make this work?
Would the HTTP Artifact Binding do the trick?
NOTE: This question should perhaps better be asked on https://security.stackexchange.com/.
If you cannot change the thick client, then you cannot use SAML, or any form of browser-based single sign-on (SSO). It's as simple as that.
Also, your approach of expecting users to type their SSO credentials into the thick client, and then sending them through HTTP basic auth and automatically typing them into a form is insecure for several reasons:
The user's plain text password needs to be passed through several instances that should never see it, and it will probably also be stored in the thick client. Even if it is encrypted (in transit or at rest), this is less secure than the password being only stored as a hash on the IdP (and in the memory or password manager of the user).
Expecting users to enter their SSO credentials anywhere else than the login form of the IdP promotes dangerous use of credentials which can make users more susceptible to phishing.
If you are planning to use the IdP for SSO to more than one client than this one legacy thick client, you can use something like application passwords:
The user logs in to a web app using SSO (SAML)
The user creates an application password (and also has the option to revoke it)
The application password is then used with the thick client.
The "application password" can be a combination of a simple unique ID (which is entered as the "username") and a long random-generated string (which is sent as the "password"). If the thick client can store these credentials, then this approach would be somewhat user-friendly, although not as secure as true SSO.
In the long run, do consider updating the thick client.
I would like to develop RESTful API for my web application. Client have to be clear JS + HTML and user have to somehow authenticate in system. I've read few articles about security and RESTful API but some point still are not clear for me. As I've understood at the first step user have to provide his credentials to server. After that server have to validate user credentials and if they are valid, sent some encoded token to user (assume it will be [user key]:[user IP]:[token creation time]). After user authentication client have to sent this this key with each API call. That's how I've understood RESTful API authentication principes.
With such solution I assume that token can be stolen and another user can access to secured user data (even if IP is included to access token and there will be validation on each request). For this purpose I plan to include token creation time but as I understand with such solution I have to renew access token each time when it expiring - in this case it's not clear for me how to implement "remember me" functionality.
It's not clear for me how to implement 100% safe authentication for my RESTful API. Maybe I'm missing something. Or maybe my understanding of authentication principes is wrong.
It depends from what authentication scenario you are using. For example when dealing with in ASP.NET MVC + REST with Basic Authentication it will produce for you token which is in the fact Base64 encoded string '{username}:{password}'. And you are right it could be stolen, that's why for Basic Auth HTTPS is must, as token goes throw Authentication header with Basic schema.
For REST security most suitable and secure are OpenId and OAuth. Just don't reinvent wheel and use already existing standards. OAuth in compare to OpenID includes not only authentication but authorization as well. OAuth already describes all nuances with token renew and token creation time and so on.
Now practical how to implement OAuth in REST. First of all read standard. For your case read with attention Implicit Grant flow, because standard has multiple flows for different client with different trust level and security.
https://www.rfc-editor.org/rfc/rfc6749
And after that you can try some already implemented library in technological stack you are using either Java or .NET. For client it is not so important to use library in compare but for server implementation
About potential security problem read here https://www.rfc-editor.org/rfc/rfc6749#section-10.
Some think that OAuth 2.0 is less secure that OAuth 1.0, and it is also dependant from token format. Anyway access token should be passed in HTTP Header and through HTTPS as well as clientid should be stored and passed securely.
I'm trying to investigate possibility of using claims aware security in the system I'm developing. The more I read about all this stuff the more confused i get.
So i decided to describe what I know already and I would ask that You correct my statements. I got lost with all those protocols and technologies used.
Below is a simple diagram of my system. There are two services - a REST service implemented using WCF and an ASP MVC web application.
I know I need to get an STS which will be a trusted identity issuer for both of my services. I'll be using roles to differentiate certain levels of access.
The STS functionality might be accomplished by using ADFS (or is there anything more needed in addition to it) or WSO2 Identity Server.
By saying STS I mean a service which will get credentials from a client wanting to authenticate and will return a set of claims (which will beside other information contain the role assigned to the user). This set of claims will be in the form of a token.
On the diagram I marked different client types with different colors. No I'll try to describe what protocols/formats i think will be used. I am pretty confused on what is a format and what is a protocol. But let me try:
Red scenario: rich WCF client authenticates agains REST service
Request to STS will be sent using WS-trust (is there any other possibility in ADFS or WSO2?). The credentials might be in one of several forms such as X.509 certificate, password digest, kerberos, windows authentication, SAML token (this is used in federation scenarios, right?), and few other.
The answer to the client will be in a form of SWT token sent over OAuth protocol since this is the way we do it when trying to authenticate against REST services.
Does ADFS support SWT and OAuth? I couldn't find the information.
The client then send the token received from STS to the REST service. Once again this is SWT token on OAuth.
I guess as for the client code all can be easily implemented using Windows Identity Framework.
Green scenario: rich android client authenticates against REST service
All protocols/formats are the same as in previous scenario. Is there any framework which can let me easily implement this?
Blue scenario: user of web browser authenticates against ASP MVC web application
User goes to the web app's main page. The webapp detects that he isn't authenticated yet so redirects him to the sign on page on STS (the sign on page is on STS, right?).
and 3. STS authenticates the user and sends HTTP response containing SAML token and redirection to webapp. So here is HTTP used, not WS-Trusts.
Another question. In this web browser scenario there will be a cookie written on a client's machine. So whenever client will try to authenticate once again, he'll just send the cookie to sts to obtain the token. There will be no need to send the credentials. STS will issue a token basing on the cookie without any real authentication logic involved. Is that statement correct?
ADFS / WIF out the box only supports SAML tokens - no OAuth support.
ADFS / WCF uses WS-Trust.
The sign-on page is part of ADFS.
The answer is "Yes" to your last question but (at some point) it will expire and the user will have to authenticate again.
Update:
Have a look at Claims Based Identity & Access Control Guide