I'm considering a microservice architecture and I'm struggle with authorization and authentication. I found a lot of resources about oauth2 and openid connect that claim they solve the issue but it is not clear enough for me.
Let's consider we have a following architecture:
In my system I want to add a feature only for a certain group of users defined by role. I want to also know the name of the user, their email and id.
After my research I find the following solution to be a good start:
SPA application displays login form.
User fills in the form and sends POST request to authN&authZ server.
The server replies with access token (being a JWT) that contains name, email, id and role of the user. The response contains a refresh token as well.
SPA application stores the token and attaches it to every request it makes.
Microservice 1 and Microservice 2 check if the token is valid. If so, they check if the role is correct. If so, they take user info and process the request.
How far away from the good solution I am? The login flow looks like Implicit flow with form post described here but with implicit consents and I'm not sure if it's fine.
Moving forward, I find passing user data in JWT (such as name, email) to be not a good solution as it exposes sensitive data. I found resources that say it is recommended to expose only a reference to a user in token (such as ID) and replace such token with a classic access_token in reverser-proxy/api gateway when sending a request to a microservice. Considering such solution I think that following scenario is a good start:
SPA application displays login form.
User fills in the form and sends POST request to authN&authZ server.
The server replies with access token and refresh token. API gateway (in middle) replaces access token with ID token and stores claims from access token within its cache.
SPA application stores the token and attaches it to every request it makes.
Handling a request, API Gateway takes ID Token and based on the user ID generates a new access token. The access token is send to microservice 1 or microservice 2 that validate it as previous.
How do you find such solutions? Is this a secure approach? What should I improve proposed flow?
Thanks in advance!
You are on the right tracks:
ZERO TRUST
This is an emerging trend, where each microservice validates a JWT using a library - see this article. JWT validation is fast and designed to scale.
CONFIDENTIAL TOKENS FOR CLIENTS
Internet clients should not be able to read claims that APIs use. The swapping tokens in a gateway concept is correct, but the usual approach is to issue opaque access tokens here, rather than using ID tokens. At Curity we call this the Phantom Token Approach.
SECURE COOKIES IN THE BROWSER
One area to be careful about is using tokens in the browser. These days SameSite=strict HTTP Only cookies are preferred. This requires a more complex flow though. See the SPA Best Practices for some recommendations on security.
SPAs should use the code flow by the way - aim to avoid the implicit flow, since it can leak tokens in the browser history or web server logs.
SUMMARY
All of the above are general security design patterns to aim for, regardless of your Authorization Server, though of course it is common to get there one step at a time.
Don't use your own login form. As Garry Archer wrote, use the auth code flow with PKCE which is the recomended flow for applications running in a browser.
If you don't want to get an ID token, don't ask for the openid scope in the initial auth request. The type of issued access tokens (JWT or opaque) can often be configured on your OAuth2 server. So I see no need to issue new tokens at your gateway. Having more token issuers opens more ways of attacking your system.
Your backend modules can use the userinfo endpoint, which will give them info about the user and validate the token. This way, if the token was invalidated (e.g. user logged out), the request processing will not proceed. If you validate just a JWT signature, you will not know about the token being invalidated.
If you plan to make requests between your backend modules as part of of a user request processing, you can use the original access token received from your SPA as long as your modules are in a safe environment (e.g. one Kubernates).
Related
I'm looking for guidance and/or best practices in implementing step-up authentication. The generic scenario is as follows: user logs in to web app using some identity provider, user then goes to some specific area of web site which needs to be protected by additional MFA, for example OTP. All functionality for the website is via REST API, authenticating with JWT bearer token.
The best description of the flow I found is from Auth0 here. Basically, user acquires the access token with scope which is just enough to access APIs that do not require additional protection. When there is a need to access secure API the authorization handler on backend would check if the token has the scope indicating that the user has completed the additional MFA check, otherwise it's just HTTP 401.
Some sources, including the one from Auth0, suggest using amr claim as an indication of passed MFA check. That means that identity provider must be able to return this claim in response to access token request with acr_values parameter.
Now, the detail that is bugging me: should the frontend know in advance the list of API that might require MFA and request the elevated permissions beforehand, or should frontend treat the HTTP 401 response from backend as a signal to request elevated permissions and try again?
Should identity provider generate relatively additional short-lived token to access restricted APIs? Then, if frontend has 2 tokens it must definitely know which token to use with which API endpoint. Or maybe identity provider can re-issue the access token with normal lifespan but elevated permissions? Sounds less secure then the first approach.
Finally, do I understand the whole process right or not? Is there some well documented and time-tested flow at all?
CLIENTS
Clients can be made aware of authentication strength via ID tokens. Note that a client should never read access tokens - ideally they should use an opaque / reference token format. Most commonly the acr claim from the ID token is used.
This can be a little ugly and a better option can sometimes be to ask the API, eg 'can I make a payment'? The client sends an access token and gets a JSON response tailored to what the UI needs. The API's job is to serve the client after all.
APIs
APIs receive JWT access tokens containing scopes and claims. Often this occurs after an API gateway swaps an opaque token for a JWT.
The usual technique is that some logic in the Authorization Server omits high privilege ones, eg a payment scope or claim, unless strong authentication was used.
It is worth mentioning plain old API error responses here. Requests with an insufficient scope typically return 403, though a useful JSON error code in an error object can be useful for giving the client a more precise reason.
STEP UP
As you indicate, this involves the client running a new code flow, with different scope / claims and potentially also with an acr_values parameter. The Curity MFA Approaches article has some notes on this.
It should not be overused. The classic use case is for payments in banking scenarios. If the initial sign in / delegation used a password, the step up request might ask for a second factor, eg a PIN, but should not re-prompt for the password.
CLIENTS AND ACCESS TOKENS
It is awkward for a client to use multiple access tokens. If the original one had scope read then the stepped up one could use scope read write payment and completely replace the original one.
In this case you do not want the high privilege scope to hang around for long. This is yet another reason to use short lived access tokens (~ 15 minutes).
Interestingly also, different scopes the user has consented to can have different times to live, so that refreshed access tokens drop the payment scope.
ADVANCED EXAMPLE
Out of interest, here is an interesting but complicated article on payment workflows in Open Banking. A second OIDC redirect is used to get a payment scope after normal authentication but also to record consent in an audited manner. In a normal app this would be overkill however.
If I authenticate with OpenID connect, I can authenticate my SPA ok.
How do I use the obtained access token now to access my own REST resources?
It's a simple question, but I don't find satisfactory answers.
A prominent answer I always find is 'use oidc when you don't have a backend'.
Now that makes me wonder if ever a webapp was created that didn't need a backend.
Oidc is almost always the answer when the question of storing a refresh token in the client pops up (like in 'use oidc, it's a better architecture and ditch the refresh token') but it doesn't really explain anything.
So when the user logs in with, say Google, he obtains an identity and an access token (to ensure that the user is who he claims he is).
So how do you use this to authenticate at your own REST service?
The only real way I see it as stateless is by sending another request at the server to the provider on every request to the REST api, to match the identity to the validity of the access token there.
If not, we fall back to the good 'ol session vs jwt discussion, which doesn't quite seem to click with the oidc because now we're duplicating authentication logic.
And the good 'ol refresh token in the browser is generally promoted as a bad idea, although you can keep access tokens in the browser session storage (according to the js oidc client library), autorefresh them with the provider and that's fine then (-.-).
I'm running again circles.
Anybody can lay this out for me and please break the loop?
Your SPA (frontend) needs to add an authorization header with access token to each API request. Frontend should implement the authorization code flow + PKCE (implicit flow is not recommended anymore) + it needs to refresh access token.
Your API (backend) needs to implement OIDC (or you can use "oidc auth" proxy in front of backend) - it just validates access token, eventually returns 401 (Unauthorized) for request with invalid/expired/... token. This token validation is stateless, because it needs only public key(s) to verify token signature + current timestamp. Public keys are usually downloaded when backends is starting from OIDC discovery URL, so they doesn't need to be redownloaded during every backend request.
BTW: refresh token in the browser is bad idea, because refresh token is equivalent of your own credentials
For the past 10+ days I've read an watched ALL the content I could find on understanding OAuth2 and OpenID Connect, only to find that many people disagree on the implementation, which really confuses me.
To my understanding, all the articles and examples I found assume you want access to eg. google calendar, profile info or emails if you eg. login with google, but I do NOT need to access other than my own API's - I only want to use Google, Facebook etc for logging in, and getting an id which I can link to my user in my own database - nothing more than that.
I'll try illustrate my use case and use that as an example.
A note on the diagram: the Authentication service could probably be built into the API Gateway - not that i matters for this example, since this is not about "where to do it", but "how to do it the best way" possible, for an architecture such as mine, where it's used for my own API's / Microservices, and not accessing Google, Facebook etc. external API's
If you can understand what I'm trying to illustrate with this diagram above, please tell me if I've misunderstood this.
The most basic requirements for this architecture you see here are:
Users can login with Google, Facebook, etc.
The same login will be used for all micro-services
OpenId user will have a linked account in the database
User access is defined in my own db, based on groups, roles and permissions
I do not intend to use external API's after the user is authenticated and logged in. No need for ever accessing a users calendar, email etc. so I really just need the authentication part and nothing else (proof of successful login). All user access is defined in my own database.
So a few fundamental questions comes to mind.
First of all, is OpenID Connect even the right tool for the job for authentication only (I'll have no use for authorization, since I will not need read/write access to google / facebook API's other than getting the ID from authenticating)?
People generally do not agree on whether to use the ID or Access token for accessing your own API's. As far as I understand the ID token is for the client (user-agent) only, and the access token is for eg. accessing google calendar, emails etc.... External API's of the OpenID Provider... but since I'll only be accessing my own API's, do I event need the access token or the ID token - what is the correct way to protect your own API's?
If the ID token is really just for the client, so it can show eg. currently logged in user, without going to the DB, I have 0 use for it, since I'll probably query the user from from the db and store it in redux for my react frontend app.
Dilemma: To store user details, groups, roles and permission inside JWT or not for API authorization?
By only storing the user identifier in the token, it means that I always allow authenticated users that has a valid token, to call endpoints BEFORE authorization and first then determine access based on the db query result and the permissions in my own database.
By storing more data about the user inside the JWT, it means that in some cases, I'd be able to do the authorization / access (group, role, permission) check before hitting the API - only possible with user info, groups, roles and permission stored inside a JWT issued upon login. In some cases it would not be possible due to eg. the CMS content access permissions being on a per-node level. But still it would mean a little better performance.
As you can see on the diagram I'm sending all API requests through the gateway, which will (in itself or with an authentication service) translate the opaque access token into some JWT with an identifier, so I can identify the user in the graph database - and then verify if the user has the required groups, roles and permissions - not from an external API, but from my own database like you see on the diagram.
This seems like a lot of work on every request, even if the services can share the JWT in case multiple services should need to cross call each other.
The advantage of always looking up the user, and his permissions in the db, is naturally that the moment the user access levels change, he is denied/granted access immediately and it will always be in sync. If I store the user details, groups, roles and permission inside a JWT and persist that in the client localstorage, I guess it could pose a security issue right, and it would be pretty hard to update the user info, groups, roles and permissions inside that JWT?
One big advantage of storing user access levels and info inside the JWT is of course that in many cases I'd be able to block the user from calling certain API's, instead of having to determine access after a db lookup.
So the whole token translation thing means increased security at the cost of performance, but is is generally recommended and worth it? Or is it safe enough to store user info and groups, roles, permissions inside the JWT?
If yes, do I store all that information from my own DB in the ID Token, Access token or a 3rd token - what token is sent to the API and determines if the user should be granted access to a given resource based on his permissions in the db? Do I really need an access token if I don't need to interact with the ID providers API? Or do I store and append all my groups, roles, permissions inside the ID token (that doesn't seem clean to me) issued by OpenID connect, and call the API and authorize my own API endpoints using that, even if some say you should never use the ID token to access an API? Or do I create a new JWT to store all the info fetched from my database, which is to be used for deciding if the user can access a given resource / API endpoint?
Please do not just link to general specs or general info, since I've already read it all - I just failed to understand how to apply all that info to my actual use case (the diagram above). Try to please be as concrete as possible.
Made another attempt to try and simply the flow:
The following answer does only apply for a OpenID Connect authentication flow with a 3rd party IDP (like Google). It does not apply for an architecture where you host your own IDP.
(There are some API gateways (e.g Tyk or Kong) which support OpenID Connect out of the box.)
You can use JWTs (ID token) to secure your APIs. However, this has one disadvantage. JWTs cannot be revoked easily.
I would not recommend this. Instead you should implement an OAuth2 authorization server which issues access tokens for your API. (In this case, you have two OAuth2 flows. One for authentication and one for authorization. The ID and access token from the IDP are used only for authentication.)
The following picture shows a setup where the API gateway and authentication/authorization server are two separate services. (As mentioned above, the authentication/authorization can also be done by the API gateway.)
The authentication flow (Authorization Code Grant) calls are marked blue. The authorization flow (Implicit Grant) calls are marked green.
1: Your web app is loaded from the app server.
2a: The user clicks on your login button, your web app builds the authorization URL and opens it. (See: Authorization Request)
2b: Because the user hasn't authenticated and has no valid session with your authorization server, the URL he wanted to access is stored and your authorization server responds with a redirect to its login page.
3: The login page is loaded from your authorization server.
4a: The user clicks on "Login with ...".
4b: Your authorization server builds the IDP authorization URL and responds with a redirect to it. (See: Authentication Request)
5a: The IDP authorization URL is opend.
5b: Because the user hasn't authenticated and has no valid session with the IDP, the URL he wanted to access is stored and the IDP responds with a redirect to its login page.
6: The login page is loaded from the IDP.
7a: The user fills in his credentials and clicks on the login button.
7b: The IDP checks the credentials, creates a new session and responds with a redirect to the stored URL.
8a: The IDP authorization URL is opend again.
(The approval steps are ignored here for simplicity.)
8b: The IDP creates an authorization and responds with a redirect to the callback URL of your authorization server. (See: Authentication Response)
9a: The callback URL is opened.
9b: Your authorization server extracts the authorization code from the callback URL.
10a: Your authorization server calls the IDP's token endpoint, gets an ID and access token and validates the data in the ID token. (See: Token Request)
(10b: Your authorization server calls the IDP's user info endpoint if some needed claims aren't available in the ID token.)
11a/b: Your authorization server queries/creates the user in your service/DB, creates a new session and responds with a redirect to the stored URL.
12a: The authorization URL is opend again.
(The approval steps are ignored here for simplicity.)
12b/+13a/b: Your authorization server creates/gets the authorization (creates access token) and responds with a redirect to the callback URL of your web app. (See: Access Token Response)
14a: The callback URL is opened.
14b: Your web app extracts the access token from the callback URL.
15: Your web app makes an API call.
16/17/18: The API gateway checks the access token, exchanges the access token with an JWT (which contains user infos, ...) and forwards the call.
A setup where the authorization server calls the API gateway is also possible. In this case, after the authorization is done, the authorization server passes the access token and JWT to the API gateway. Here, however, everytime the user infos change the authorization server has to "inform" the API gateway.
This is a very long question. But I believe most can be summarised by answering below,
To my understanding, all the articles and examples I found assume you want access to eg. google calendar, profile info or emails if you eg. login with google,
You do not necessarily use Access token (ID token in some occasions) to access the services offered by token issuer.You can consume tokens by your own APIs. What these Identity Providers (synonym to Authorization server, or IDP in shorthand) is to hold identities of end users. For example, typical internet have a Facebook account. With OAuth and OpenID Connect, the same user get the ability to consume your API or any OAuth/OIDC accepted service. This reduce user profile creation for end users.
In corporate domain, OAuth and OIDC serves the same purpose. Having a single Azure AD account lets you to consume MS Word as well as Azure AD's OIDC will issue tokens which can be used to Authorise against an in-house API or an third party ERP product (used in organization) which support OIDC based authentication. Hope it's clear now
A note on the diagram is that the Authentication service could probably be built into the API Gateway - not sure if that would be better?
If you are planning to implement an API gateway, think twice. If things are small scale and if you think you can maintain it, then go ahead. But consider about API managers which could provide most of your required functionalities. I welcome you to read this article about WSO2 API manger and understand its capabilities (No I'm not working for them).
For example, that API manager has built in authentication handling mechanism for OAuth and OIDC. It can handle API authentication with simple set of configurations. With such solution you get rid of the requirement of implement everything.
What if you can't use an API manager and has to do it yourself
OpenID Connect is for authentication. Your application can validate the id token and authenticate end user. To access APIs through API Gateway, I think you should utilise Access token.
To validate the access token, you can use introspection endpoint of the identity provider. And to get user information, you can use user-info endpoint.
Once access token is validated, API gateway could create a session for a limited time (ideally to be less or equal to access token lifetime). Consequent requests should come with this session to accept by API gateway. Alternatively, you can still use validated access token. Since you validated it at the first call, you may cache for a certain time period thus avoiding round trips to validations.
To validate user details, permission and other grants, well you must wither bind user to a session or else associate user to access token from API gateway at token validation. I'm also not super clear about this as I have no idea on how your DB logic works.
First Appreciate your patience in writing a very valuable question in this forum
we too have same situation and problem
I want to go through ,as images are blocked in our company in detail
Was trying to draw paralles to similar one quoted in the book
Advance API In Practise - Prabath Siriwerdena [ page 269]Federating access to API's Chapter. Definitely worth reading of his works
API GW should invoke Token Exchange OAUTH2.0 Profile to IDP [ provided the IDP should support TOken Exchange profile for OAUTH 2.0
The Absence of API Gateway , will result in quite bespoke development
for Access Control for each API Check
you land up in having this check at each of the api or microservice [ either as library which does work for you as a reusable code]
definitely will become a choking point.]
Watching this video, it details in OAuth2 that the client application first has to get the authorization grant from the Authorization server and then use that grant to get a token before being able to access the resource server. What purpose does the grant serve? Why not give the client the token right away after the user signs on with his/her username and password?
Because it is more secure, for some application types.
What you describe is so called authorization-code-flow. It is normally used for "classical" web applications, where only the backend needs to access resource server. The exchange of authorization code to access token happens on the backend and access token never leaves it. Exchange can be done only once and in addition client id and secret (stored on the backend) are necessary.
Single-Page-Applications often use implicit-flow where access token is delivered to the frontend directly in the URL.
See more here:
IdentityServer Flows
EDIT: Q: "I still don't see how it is more secure given that you have to have the grant in order to get the token. Why need 2 things instead of just 1 thing to access the resource? If someone steals the token, they can access the resource anyway – stackjlei"
"Stealing" access token will work independent on how your application acquires it. However, stealing access token on the backend is much more difficult than on the frontend.
Authorization code is delivered to the backend also over the frontend but the risk that someone intercepts and uses it is tiny:
It can be exchanged only once.
You need client-id and client-secret in order to exchange it. Client-secret is only available on the backend.
Normally, authorization code will be exchanged by your backend to access-token immediately. So the lifetime of it is just several seconds. It does not matter if someone gets hold of used authorization code afterwards.
In your scenario there could be two servers, an Authorization and a Resource one.
It could be only one as well, but let's imagine this scenario.
The purpose of the Authorization Server is to issue short lived access tokens to known clients. The clients identify themselves via their CLientID and CLientSecret.
The Authorization Server ( AS ) holds the list of clients and their secrets and first checks to make sure the passed values match its list. If they do, it issues a short lived token.
Then the client can talk to the Resource Server ( RS ), while the token is valid. Once the token expires, a new one can be requested or the expired one can be refreshed if that is allowed by the Authorization Server.
The whole point here is security, Normally, the access tokens are passed in the Authorization header of the request and that request needs to be over https to make sure that the data can't be stolen. If, somehow, someone gets hold of an access token, they can only use it until it expires, hence why the short life of the tokens is actually very important. That's why you don't issue one token which never expires.
You have different type of OAuth. On type doesn't require to use the 'grant' authorization. It depend who are the user/application, the ressource owner and the server API.
This way, you - as a user - don't send the password to the application. The application will only use the grant token to gain access to your ressources.
I think this tuto is a pretty good thing if you want more details
https://www.digitalocean.com/community/tutorials/an-introduction-to-oauth-2
The specification of OAuth2 states that an authorization server must not issue a refresh token when using implicit grant. In our use case we protect a RESTful API with OAuth2 and use a Single Page Javascript application as a client for this API. As it would be very difficult to redirect to the authorization server after an access token has expired, we are searching for a better way to get a new valid token. I could think about two different approaches and wonder which one could be better:
Use a hidden iframe to Rerequest a valid access token. For this it is necessary to include a parameter like “prompt=none” which tells the OAuth provider neither to challenge authentication, nor to display an authorization page. If the user is authenticated and has authorized the application the server will send back an access token in the urls # parameters. If one of the previous conditions is not fulfilled, it will redirect with an error like #error=authentication%20lost. With this behaviour we can use short lived access tokens also with an implicit flow.
We could use an additional scope (e.g. offline) which tells the server to hand out a refresh token. Even if the original spec says that implicit flow does not issue refresh tokens (which is correct if the client only uses OAuth it for a first authorization) you are free to define your own scopes for your particular application. You should consider to only allow this scope from well-known clients.
Both approaches are very similar to those of OpenID Connect. Unfortunately there are not many implementations of OpenID Connect at the moment. So first step would be to extend the OAuth2 server until OIC will be more popular.
So which approach should be preferred?
EDIT: The token endpoint needs client authentication, which is only possible for confidential clients like server-side applications. With the second approach it would only be possible to let the RESTful API in our case the resource provider to refresh the token and send it back to the client. I think this would be a security risk. So probably we have only one valid approach.
I'm trying to achieve the exact same thing at the moment.
I've actually implemented hidden iframe approach and then realized you have to be very careful with iframes. Any malicious website can contain your iframe and get access token easily if you don't specify X-Frame-Options.
Best approach for refreshing token should be password grant as specified by the spec. (I wanted my users to login with their facebook account & implicit flow was easier to develop this. I have not quite figured out how to do this with password grant.)
2nd approach also came accross my mind and seems much safer than the 1st to me, since you can usually trust the https & browser storage to keep your tokens secret.
Edit
I realized, even with X-Frame-Options most browsers can't prevent redirects, because this header is attached to the response body and redirected URL will be exposed, therefore access tokens exposed.
Update
Looks like hash fragment is protected by the browser when accessed from the parent page within different domain. So I assume #access_token is safe. My bad. Just as a reminder callback page has to store the access token in its own right, instead of (my original intention) delegating it to the parent page like window.parent.storeAccessToken(hash); which obviously is a dumb thing to do.
From the OAuth0 website:
If you need to authenticate your users without a login page (for example, when the user is already logged in via SSO scenario) or get a new access_token (thus simulate refreshing an expired token), you can use Silent Authentication.
As for the Silent Authentication:
However, redirecting users away from your application is usually considered disruptive and should be avoided, from a UX perspective. Silent authentication lets you perform an authentication flow where Auth0 will only reply with redirects, and never with a login page.
This will allow you to log back the user using an SSO token, without having to prompt him for credentials again.