I'm working on a personal project composed of an API and 4 clients (web, android, iOS, windows phone).
I'm using django-rest-framework and oauth2 toolkit on the API side and I wonder which grant_type would be more suitable in my situation.
I read somewhere that the implicit grant_type is appropriate for working with mobile clients.
I'm currently using the resource owner password credentials system.
My current workflow is:
The user creates an account on the API registration page (http://mysite/api/register) then gets redirected on the web client.
The user have to authenticate himself on the API from the web client (the secret and client ID are store in the web client). If the authentication is successful the access_token and refresh_token are both stored in the user session.
Each time the user want to access a page I verify if he is authenticated by requesting the API using his access_token. If the request fails, I retry with the refresh_token. If it's fails again I redirect the user on the auth page.
The user can use the API on a mobile client with the same account without extra manipulations (the secret and client ID are store in a secure location ex. share preferences or keychain)
I like this workflow, it's simple and convenient for the user: he registers once and can use all the clients and I get a perfect separation between the logic (API) and the UI (client). But I'm worried about the security of this system. I don't want to expose my users to threats. Do you guys have any thoughts, recommendations, suggestions?
You help in this matters would be very appreciated.
Thanks in advance!
Related
I am building an ASP.NET Core 6 Web API application for mobile clients (and maybe later SPA JS app). The application should have sign-in with Google option. I also want to add my own app's custom sign up and sign in options that would also be based on JWT authentication and not cookie.
I understand that for my custom sign in flow my app will generated JWT that will be sent to the client.
But I have few questions how that works when user signs-in with its Google account:
who's responsibility is to generate the JWT when user signs-in with its Google account? Is that responsibility of Google or mine application? I don't want Google to return JWT to the client in the cookie.
Then when client is authenticated with Google, and sends requests to my application, how can my application validate JWT token it gets?
When user signs in with Google for the first time, should I automatically register that user in my application (I am using Identity framework) by taking claim values (email) from the JWT? What is the general practice here?
I am trying to understand these processes and flows so sample code is not necessary (but I do welcome it).
Ad.1. Normally, in a larger system, you would have an authorization server (AS) that would handle user authentication and the issuance of tokens. Your clients would contact only the AS, and the AS will be able to provide the user with different forms of authentication: e.g., through your website's password or through Google. The AS is the single point of issuing tokens to your clients. It can issue tokens regardless of the authentication method used. So it then doesn't matter whether the user authenticated with Google or a password, the client will still get the same access token.
Ad.2. When the AS issues token to your client, then you don't have any problems validating that token. The client doesn't care if the user authenticated with Google or not, it's not relevant in this case.
If you decide to skip using an AS and let the client receive tokens directly from Google, then you can still verify them. An ID token is a JWT and can be easily validated with a JWT library using verification keys provided by Google. Access tokens returned by Google are opaque tokens (If I remember correctly), and you need to check whether Google exposes an endpoint to verify them.
Ad.3. That is the general practice. When the user authenticates with Google and you notice that you don't have that user's data in your system, then you take the information from Google's ID token and create a user entry in your system.
I'm using an external service called auth0 in order to get an access token and let my users use my api. Auth0 is using Oauth2 protocol.
In short The user adds a username and a password, I'm doing a call to auth0 by using a client_id (apps have an id) and client_secret and I get an jwt access token in return. Then from there I carry this access token to have access to my own api since I can check its validity.
I have been looking around about how secure it is to store client_id and client_secret on the client side (e.g. web (javascript)/mobile (native or hybrid with ionic)) and everybody was saying that it's not secure since everybody can reverse engineer the code and get the client_id and client_secret. Ok...I can take it...what Can I do with them if I don't have credentials in order to get the access token?
Given that I don't want to store the client_id and the client_secret, one solutions I have thought is to make a direct call to my api (Java) with the credentials and then my api make a call to auth0 and return the corresponding access token. In this way the client_id and client_secret is stored in the backend and somebody cannot get them easily. Is that safe?
However I have some endpoints, e.g. creating use account, sending sms for phone validation etc, that cannot have credentials. How do I protect the api in such case? If I can't store my own access token on the client side how could I get an access token and access my own api without credentials?
Thanks
One possible solution that OAuth spec suggests is that you could have three different servers for your application.
client-side
backend server and an additional authentication server.
The preferred way of doing this would be that the client would send the user credentials to the authentication server. The authentication server would be a back-end server which contains the client secret
The authentication server will authenticate the credentials and return back the token.
The client will then use the token obtained from the authentication server to access the resource API server.
If you wanna know more check out this video
https://www.youtube.com/watch?v=rCkDE2me_qk
In my opinion you are almost certainly using the wrong OAuth flow. I use Auth0 with Ionic as both a web app and a native Cordova app. I don't have the client secret in my client code at all.
If you follow the Auth0 quickstarts (https://auth0.com/docs/quickstarts), you should be choosing (Native/Mobile App) if you are deploying to app stores, and (Single-Page App) if you are deploying the web version of Ionic. From there you can pick Cordova (for native) or Angular (for SPA). These should give you instructions that implement OAuth flows which DO NOT require your client secret. My guess would be you are referencing a "Regular Web App" quickstart, which runs server-side and CAN safely hold the client secret. That's not the world you're coding in if you are using Ionic Hybrid/Native.
I would consider wrapping the call to Auth0 into your own server side implementation as safe. Your API takes user credentials and then calls Auth0 and this way your client_id/secret are secure on your server and the client can be reverse-engineered all the way without compromising your security.
Regarding the other APIs which cannot have credentials you are pretty much out of luck. Their very use case is to be used by an unauthenticated third party, so at least the account creation API cannot really be protected. However you can still use some nicely designed constraints to limit the attack surface. E.g. you can require an email address/phone number to register and you will not allow the same address/phone number twice. If you set up your process that you first need to confirm your email address before you can validate your phone number this will make the life of an attacker a lot harder. He would need a real working email address, and some automation to receive your confirmation mails before he could get to call your SMS service. You could also rate-limit the service per IP-address so an attacker cannot cause your SMS cost to skyrocket by issuing a lot of calls for SMS validation in a short period of time.
I have a question more related to the way OAuth 2 is working but since using IdentityServer to implement OAuth I think it's relevant. I could not find an answer anywhere.
I'm building a website and a mobile app that consumes my own API. Each user of my app will have a username and password, that will give him access to the app/website and though the API to his information.
I'm not sure about the right way to handle the flow for user login:
On the website I have my own designed login form. I don't want to move the user to my auth server to login, and then have him approve the information he gives - he is the user on my system - I have access to all information - kida like facebook has a login and access to the informatio - they don't ask what you're willing to give them. So is implicit really the way for this?
On the mobile app I also have a login form and now I read here (https://datatracker.ietf.org/doc/html/draft-ietf-oauth-native-apps-10) that the OAuth approach is to have the login in a WebView?? Doesn't look like facebook login is in a WebView on their mobile app.
The approach I was first lookin at is the Resource Owner. Users will login, get the token and the refresh token and can start working against my APIs. But storing my client_id and secret on the mobile app? on the website javascript files? doesn't feel right. I can of course make a call to an API that will mask those and be a proxy to the login process... but... (read #4).
In the future I would like to allow access for third-party developers. For them to allow login for users of my system I will use the implicit flow. Also, I plan for those developer accounts to have restricted API access (for example, the number of calls to the API will be limited by plan). What prevents those developers from asking for the username and password of their account on my system on their website, getting the response from my servers with the access token and refresh token, and using my API however they want, without restrictions, and having access to the entire user profile?
Lets say I'm sticking to the resource owner flow, receiving back from the server a token and a refresh token. What should I store on the mobile device and how? What should be stored in the browser and how? the refresh token? and each time he opens the app get a new updated token with that refresh token?
Edit
Just to clarify, because I find a lot of lectures and articles that explain the process from an API consumer point of view (ie. the third-party developer): I am the API owner and the auth server owner, I'm the owner of the user accounts (they are my users of my services), I'm also my own consumer (though the website and the mobile app), and in the future I want to enable third-party developers to allow my users to login with their accounts of my service (kinda like Facebook or Google)
You're correct that you shouldn't store the client_secret in your app, but I doubt you will get around storing the client_id. You could disable the consent screen for your app as well, and build a native login view. You need to store the access_token and the refresh_token on the device (maybe encrypted in a database) if you don't want the user to login everytime they use your app.
As for problem 4, you could do the following:
Embed the client_secret in your (web) app
Set up which hosts have access to your api on the IdentityServer
The IdentityServer generates a salt and sends it to the client
The client calculates a session_secret using hash(ip_address + session_salt)
The client uses the session_secret and the client_secret for the API call
Server validates the hash and client_secret
It's nearly impossible to completely prevent someone from using your API. But you should add various rate limiting methods, such as limiting IP addresses, API calls etc. But nothing will stop someone decompiling your app and accessing your client_id.
I am currently being tasked with writing an RESTful API that we are going to release to our clients. I plan to use Web API 2 with OWIN middleware. The applications that will be accessing this can range anywhere from a Console app with no additional user authentication to full blow web applications that will involve user login and authentication.
One other requirement is that we need not only provide user authentication to control what data we will return, we also need to authenticate that the application accessing our API is allowed by our organization. So this requires us to not only validate the application invoking our API's but also the many users logging in via that application.
The goal is to use a bearer token technology like OAuth 2. I am not very familiar with OAuth 2 but after looking it over I don't see a way to facilitate a two part authentication like this. Ideally we would like to issue a clientId and a clientPassword and they would pass this in along with the userid and userpassword and we would then pass a token back.
Is this something that anyone has tackled before using OAuth2 or is this a job for a custom auth solution?
What you describe sounds like fairly standard Oauth2.0 scenarios so you shouldn't need anything custom.
For the web apps you could use the Resource Owners Password Credentials Grant. Which is a standard flow.
This allows your Oauth client to pass in its client Id and client secret along with a user's Id and password to your Auth Server which can then validate both the user and the client before issuing a token in response.
But I would recommend using the Authorization Code Grant which is probably the best option for web applications and it involves both the client and the user being effectively authenticated.
For you console apps, or any apps which are not user specific, you can use the Client Credentials Grant. This allows the app to authenticate and to get a token without a user.
If you need your API to be aware of the client application, the only flow you can't really use is the Implicit Grant which doesn't authenticate the client.
I'm developing a common auth service for several different web services. The general idea is that a user goes to one of these web services and clicks on a login button and gets redirected to my auth service. Once authenticated, the auth service redirects the user back to the original web service.
To begin with, the auth service will use Google's OpenID Connect service. One idea I had was that when a user was authenticated using Google I could pass around the resulting JWT to other services to use as an auth token. Google mentions this idea in their documentation (https://developers.google.com/accounts/docs/OAuth2Login#validatinganidtoken):
One thing that makes ID tokens useful is that fact that you can pass them around different components of your app. This can be helpful because different components of your application may pass the ID token between them and use it as a lightweight authentication mechanism authenticating the app and the user.
Can anyone help me with the actual details on how this should work e.g. How do I handle log outs both on Google and from the auth service? There appeared to be some documentation on this in older OpenID Connect Session documents (-03) but it appears to have been removed.
You should not pass the ID Token that you receive to Google back to the app to use as an auth token. The ID Token will have an audience specific to your app, not the ultimate client.
If you just passed this back to clientA, then you would open it up to auth attacks where clientB could obtain an Google ID Token via your auth service and then use that token to gain access to user data in clientA.
A much better approach would be to mint your own ID Token and provide an audience designation for each app that uses your service.
Monitoring session state is not as straightforward. The easiest thing to do is for your service to wrap G+ sign-in libraries and interface with the session state api's.
https://developers.google.com/+/web/signin/session-state