Currently I'm developing a GAE application in golang that would allow the users to log into the service using OpenID from Google. Each account would be mapped to one Google account obviously. Should I implement captcha for user registration, or does using the OpenID already prevent the service from a bot attack?
Are you accepting any OpenID? Then this is not sufficient protection. Anyone can put up an OpenID site on their own domain, use a script to create a billion accounts, and log in with one of them on your site.
Are you accepting only specific OpenID providers? Then you're not using OpenID as it was meant to be used (and, in fact, expressly the opposite of why it was created). But the answer then would be: do you trust that the OpenID providers have done their own due diligence at the time of account creation to prevent spam accounts?
Regardless, you'll likely find that captcha's a poor solution, and currently all implementations of it are broken (but that's a whole different topic).
First of all, fear of a bot attack is really a preliminary caution. Parsing CSRF tokens (which you should use anyway) and dealing with cookies would already cause enough pain to an attacker. I doubt anyone would trouble himself into setting up very own OpenID provider which is the only way one can do that.
Secondly, App Engine's OpenID Federated Login is the User API feature, not Go runtime's one — you can use any OP with it (Facebook, Yahoo!, Steam... Google). User API provides two modes (configurable via Application Settings): OpenID Federated Login and Google Accounts sign in. Which to pick depends on requirements. If all you need is user's Google account you'd better use the latter, when if it's simple password-less sign in with nearly everything, choose OpenID. From the app's perspective both look identical. Best way to go is OAuth, though.
Getting back on the question, both Google Accounts API and OAuth would make such attacks almost impossible and with OpenID it is a little simpler.
https://developers.google.com/appengine/docs/go/users/overview
Related
I have a public-facing application that uses ASP.NET Core Identity to store first-party logins, and no intention of using third-party IdPs like Facebook or Google. I want to build the front-end in React, and the application comprises an API fronting a couple of back-end services to which I'll need to forward JWTs for authorisation.
The plan so far is to use Identity Server 4 as the IdP for the project, backing it into the ASP.NET Core Identity data stores.
Current guidance is to use Authorization Code Flow with PKCE, which would require redirection to the IdP, two sets of styling etc.
In this scenario, where there is no possibility of a third-party IdP, is Resource Owner Password Grant still highly discouraged? On the face of it, it gives a neater experience:
User populates React-based login page
XHR POST to IdP with credentials (modulo an MFA challenge)
IdP returns an access token, React app subsequently uses that for future requests to the API
What issues will I introduce by pursuing the ROPC grant in this specific situation, vs accepting the need and duplication involved in a redirect-based flow to the IdP?
AMOUNT OF WORK
This is one of the big issues. As well as a login screen you'll have to make sure other areas such as Forget Password also work. If you build a second app you'll need to make it work there also.
EXTENSIBILITY
This article summarises problem areas. One of these is no scope to ever extend login solutions.
SECURITY
Token refresh does not (usually) work with ROPG, which leads to long lived access tokens and other complexity. Also, with OAuth it is recommended that the app never sees credentials.
From a security viewpoint it looks more modern to redirect the user - all the big providers do it - eg Google, Microsoft.
BRIDGING SOLUTION
As you say, if the password is private to your app it may not be the worst thing in the world. Capturing a user's Google password in your app would be a bad thing though.
ROPG has its uses but does not have much of a future - it is deprecated in OAuth 2.1 and providers will be getting rid of it. So I would also recommend what LalitaCode suggests ..
You can create a React based Identity Server login page for Authorization Code flow with PKCE instead of using MVC UI if you want. It is just extra work and complicated. I would recommend you just style the Identity Server MVC UI to look exactly like your frontend SPA. This is the simplest way and the path I took when I did a project with Identity Server(with Angular as front end).
Currently have an client application running on GAE that supports OpenId login to access resources on a server application (also on GAE) through OAuth. All is fine.
But, it requires that the user have an account with Google, Yahoo, or other OpenId provider. While that seems fine for personal users, how do I handle the corporate case where users want to use myname#mycompany.com as their login?
Do I build my own OpenId provider, host it, and store those names/passwords in a database?
Is there existing OpenId source that I can use, but give it my own names/passwords?
Do I skip OpenId entirely and somehow get my own custom db lookup to integrate with GAE authentication?
Something complete different?
My work is in Java, so Java solutions are preferred. And since this is more for prototyping than for production use, easier solutions are preferred :).
After doing some additional research it seems that if I want to stick with my current use of Restlet's GAEAuthenticator then I will have to rely on OpenId. There appears to be ways to rewrite the entire authentication so that it works both with Google/OpenId and a custom database (not what I want to try), and there appear to be commercial products (SadaSystems) that seem to also do it.
For now, my workaround is to ask people who want to use their own email address in our system to sign up with MyOpenId and add that email address as their persona.
I'm coding a Python - Google App Engine application. There are 2 important things this app must do:
Write in user's calendar.
Write in user's profile (working with users in a Google Apps domain)
First operation is easy. If I understood OAuth, this is the classical 3-legged scenario. The scenario for which OAuth was originally developed. The logged user provides credentials to the client (my app) to access the user's data (calendar) on his/her behalf. So, this can be done just with the logged user's credentials.
Second operation is not so easy. It can't be done with just the logged user's credentials. This is a 2-legged OAuth scenario. So, I need to delegate in a Google Apps domain admin's account to access the users's profiles using Google Profiles API (via google data library). So far, I hardcode admin user/password in a json file, and my app loads that file. But that sounds kind of dirty for me.
At the end, my app needs to handle the classical 3-legged OAuth scenario (no problem, just need the logged user's credentials) plus a 2-legged OAuth scenario (need administrator credentials).
Is there any official or more elegant way to handle both scenarios in Google App Engine, working with Google Apps domain, that hardcoding admin credentials?
Many thanks in advance
AFAIK there is no way to authorize a write operation to Google Contacts Data API using 2-legged OAuth.
Google's documentation specifies which APIs are accessible via 2-legged OAuth, but it seems that someone at Google's forgot to specify that some of them are read-only :-S
Some people had the same problem here and here.
I look forward someone at Google to fix this. Until then, harcoding admin-level credentials is the only option I know that it works. I don't like it at all: it's dirty but effective. If someone knows a more elegant zen-level way, please illuminate us!
Is there a way to use Google accounts to authenticate my own personal website? I understand Google provides an API for authentication but it requires redirection. Is there anyway to have a customized login process without being redirected to Google?
By the way I'm a total noob when it comes to user authentication. I did try to find a preexisting thread but was unable to find one.
For exactly the reason given by Kris, Google (along with similar OAuth providers) doesn't allow you to authenticate users directly.
It's a security measure to protect both themselves and their users from the Web equivalent of ATM card skimmers.
I have a usability answer rather than a technical one.
I am unsure if this is doable but, even if it is doable, I would highly suggest NOT doing it. Think of it this way: you want a user to use their Google account to login to your website. If you were that user, would you feel more comfortable entering your login information on a Google website or someone else's website that has nothing to do with Google?
How can I make my desktop application into an OpenID relying party?
Have a look at this question, it looks to me that OAuth can do this for you.
Since OpenID has a specified protocol, it will either work all the time (assuming providers are conforming) or not at all. Additionally, since the OpenID protocol (at least the authentication bits, see the specification) consists of simple HTTP requests, you could implement the protocol in a desktop application assuming you know how to make such requests.
It is generally considered a bad idea though, since there are better technologies (OAuth) and it disrupts the user experience (being different from OpenID in a browser — see OpenID For Desktop Applications: How? When?).
Further reading: OpenID for Desktop Clients
Don't do it.
Even an attempt to do so shows a fundamental lack of understanding in the security model that OpenID offers. You have to ask yourself what the password that you would otherwise use for your desktop app is protecting. Is it protecting assets on the local machine itself? If so, OpenID is useless because it would be a simple matter to spoof the network such that I could hack my way into the desktop app without owning the OpenID. You're wide open to user identify spoofing. Are you trying to protect network assets? OpenID fails again, since it doesn't authorize your desktop app to access those network assets, suggesting that some other authentication is going on behind OpenID so once again you're not adding any value.
OAuth is the protocol suited to allow your desktop app to access and protect network resources. If you're protecting local assets on the desktop computer, local encryption is the only way to go.
It seems to me that you should present a small web browser window for the authentication to be completed. The interaction would not be going through your code.
I would not suggest using web scrapers to do this. In doing so, you place your code in between the user's server and the user, which is a breach of the covenant that the user's password is not seen by the relying party.
OAuth is the right technology to use for a desktop application, but it doesn't use the existing password ecosystem that OpenID has, which was not a part of the OP's question.