If I get a JWT and I can decode the payload, how is that secure? Couldn't I just grab the token out of the header, decode and change the user information in the payload, and send it back with the same correct encoded secret?
I know they must be secure, but I just would really like to understand the technologies. What am I missing?
JWTs can be either signed, encrypted or both. If a token is signed, but not encrypted, everyone can read its contents, but when you don't know the private key, you can't change it. Otherwise, the receiver will notice that the signature won't match anymore.
Answer to your comment: I'm not sure if I understand your comment the right way. Just to be sure: do you know and understand digital signatures? I'll just briefly explain one variant (HMAC, which is symmetrical, but there are many others).
Let's assume Alice wants to send a JWT to Bob. They both know some shared secret. Mallory doesn't know that secret, but wants to interfere and change the JWT. To prevent that, Alice calculates Hash(payload + secret) and appends this as signature.
When receiving the message, Bob can also calculate Hash(payload + secret) to check whether the signature matches.
If however, Mallory changes something in the content, she isn't able to calculate the matching signature (which would be Hash(newContent + secret)). She doesn't know the secret and has no way of finding it out.
This means if she changes something, the signature won't match anymore, and Bob will simply not accept the JWT anymore.
Let's suppose, I send another person the message {"id":1} and sign it with Hash(content + secret). (+ is just concatenation here). I use the SHA256 Hash function, and the signature I get is: 330e7b0775561c6e95797d4dd306a150046e239986f0a1373230fda0235bda8c. Now it's your turn: play the role of Mallory and try to sign the message {"id":2}. You can't because you don't know which secret I used. If I suppose that the recipient knows the secret, he CAN calculate the signature of any message and check if it's correct.
You can go to jwt.io, paste your token and read the contents. This is jarring for a lot of people initially.
The short answer is that JWT doesn't concern itself with encryption. It cares about validation. That is to say, it can always get the answer for "Have the contents of this token been manipulated"? This means user manipulation of the JWT token is futile because the server will know and disregard the token. The server adds a signature based on the payload when issuing a token to the client. Later on it verifies the payload and matching signature.
The logical question is what is the motivation for not concerning itself with encrypted contents?
The simplest reason is because it assumes this is a solved problem for the most part. If dealing with a client like the web browser for example, you can store the JWT tokens in a cookie that is secure (is not transmitted via HTTP, only via HTTPS) and httpOnly (can't be read by Javascript) and talks to the server over an encrypted channel (HTTPS). Once you know you have a secure channel between the server and client you can securely exchange JWT or whatever else you want.
This keeps thing simple. A simple implementation makes adoption easier but it also lets each layer do what it does best (let HTTPS handle encryption).
JWT isn't meant to store sensitive data. Once the server receives the JWT token and validates it, it is free to lookup the user ID in its own database for additional information for that user (like permissions, postal address, etc). This keeps JWT small in size and avoids inadvertent information leakage because everyone knows not to keep sensitive data in JWT.
It's not too different from how cookies themselves work. Cookies often contain unencrypted payloads. If you are using HTTPS then everything is good. If you aren't then it's advisable to encrypt sensitive cookies themselves. Not doing so will mean that a man-in-the-middle attack is possible--a proxy server or ISP reads the cookies and then replays them later on pretending to be you. For similar reasons, JWT should always be exchanged over a secure layer like HTTPS.
Let's discuss from the very beginning:
JWT is a very modern, simple and secure approach which extends for Json Web Tokens. Json Web Tokens are a stateless solution for authentication. So there is no need to store any session state on the server, which of course is perfect for restful APIs.
Restful APIs should always be stateless, and the most widely used alternative to authentication with JWTs is to just store the user's log-in state on the server using sessions. But then of course does not follow the principle that says that restful APIs should be stateless and that's why solutions like JWT became popular and effective.
So now let's know how authentication actually works with Json Web Tokens. Assuming we already have a registered user in our database. So the user's client starts by making a post request with the username and the password, the application then checks if the user exists and if the password is correct, then the application will generate a unique Json Web Token for only that user.
The token is created using a secret string that is stored on a server. Next, the server then sends that JWT back to the client which will store it either in a cookie or in local storage.
Just like this, the user is authenticated and basically logged into our application without leaving any state on the server.
So the server does in fact not know which user is actually logged in, but of course, the user knows that he's logged in because he has a valid Json Web Token which is a bit like a passport to access protected parts of the application.
So again, just to make sure you got the idea. A user is logged in as soon as he gets back his unique valid Json Web Token which is not saved anywhere on the server. And so this process is therefore completely stateless.
Then, each time a user wants to access a protected route like his user profile data, for example. He sends his Json Web Token along with a request, so it's a bit like showing his passport to get access to that route.
Once the request hits the server, our app will then verify if the Json Web Token is actually valid and if the user is really who he says he is, well then the requested data will be sent to the client and if not, then there will be an error telling the user that he's not allowed to access that resource.
All this communication must happen over https, so secure encrypted Http in order to prevent that anyone can get access to passwords or Json Web Tokens. Only then we have a really secure system.
So a Json Web Token looks like left part of this screenshot which was taken from the JWT debugger at jwt.io. So essentially, it's an encoding string made up of three parts. The header, the payload and the signature Now the header is just some metadata about the token itself and the payload is the data that we can encode into the token, any data really that we want. So the more data we want to encode here the bigger the JWT. Anyway, these two parts are just plain text that will get encoded, but not encrypted.
So anyone will be able to decode them and to read them, we cannot store any sensitive data in here. But that's not a problem at all because in the third part, so in the signature, is where things really get interesting. The signature is created using the header, the payload, and the secret that is saved on the server.
And this whole process is then called signing the Json Web Token. The signing algorithm takes the header, the payload, and the secret to create a unique signature. So only this data plus the secret can create this signature, all right?
Then together with the header and the payload, these signature forms the JWT,
which then gets sent to the client.
Once the server receives a JWT to grant access to a protected route, it needs to verify it in order to determine if the user really is who he claims to be. In other words, it will verify if no one changed the header and the payload data of the token. So again, this verification step will check if no third party actually altered either the header or the payload of the Json Web Token.
So, how does this verification actually work? Well, it is actually quite straightforward. Once the JWT is received, the verification will take its header and payload, and together with the secret that is still saved on the server, basically create a test signature.
But the original signature that was generated when the JWT was first created is still in the token, right? And that's the key to this verification. Because now all we have to do is to compare the test signature with the original signature.
And if the test signature is the same as the original signature, then it means that the payload and the header have not been modified.
Because if they had been modified, then the test signature would have to be different. Therefore in this case where there has been no alteration of the data, we can then authenticate the user. And of course, if the two signatures
are actually different, well, then it means that someone tampered with the data.
Usually by trying to change the payload. But that third party manipulating the payload does of course not have access to the secret, so they cannot sign the JWT.
So the original signature will never correspond to the manipulated data.
And therefore, the verification will always fail in this case. And that's the key to making this whole system work. It's the magic that makes JWT so simple,
but also extremely powerful.
The contents in a json web token (JWT) are not inherently secure, but there is a built-in feature for verifying token authenticity. A JWT is three hashes separated by periods. The third is the signature. In a public/private key system, the issuer signs the token signature with a private key which can only be verified by its corresponding public key.
It is important to understand the distinction between issuer and verifier. The recipient of the token is responsible for verifying it.
There are two critical steps in using JWT securely in a web application: 1) send them over an encrypted channel, and 2) verify the signature immediately upon receiving it. The asymmetric nature of public key cryptography makes JWT signature verification possible. A public key verifies a JWT was signed by its matching private key. No other combination of keys can do this verification, thus preventing impersonation attempts. Follow these two steps and we can guarantee with mathematical certainty the authenticity of a JWT.
More reading: How does a public key verify a signature?
I would explain this with an example.
Say I borrowed $10 from you, then I gave you an IOU with my signature on it. I will pay you back whenever you or someone else bring this IOU back to me, I will check the signature to make sure that is mine.
I can't make sure you don't show the content of this IOU to anyone or even give it to a third person, all I care is that this IOU is signed by me, when someone shows this IOU to me and ask me to pay it.
The way how JWT works is quite the same, the server can only make sure that the token received was issued by itself.
You need other measures to make it secure, like encryption in transfer with HTTPS, making sure that the local storage storing the token is secured, setting up origins.
Ref - JWT Structure and Security
It is important to note that JWT are used for authorization and not authentication.
So a JWT will be created for you only after you have been authenticated by the server by may be specifying the credentials. Once JWT has been created for all future interactions with server JWT can be used. So JWT tells that server that this user has been authenticated, let him access the particular resource if he has the role.
Information in the payload of the JWT is visible to everyone. There can be a "Man in the Middle" attack and the contents of the JWT can be changed. So we should not pass any sensitive information like passwords in the payload. We can encrypt the payload data if we want to make it more secure. If Payload is tampered with server will recognize it.
So suppose a user has been authenticated and provided with a JWT. Generated JWT has a claim specifying role of Admin. Also the Signature is generated with
This JWT is now tampered with and suppose the
role is changed to Super Admin
Then when the server receives this token it will again generate the signature using the secret key(which only the server has) and the payload. It will not match the signature
in the JWT. So the server will know that the JWT has been tampered with.
Only JWT's privateKey, which is on your server will decrypt the encrypted JWT. Those who know the privateKey will be able to decrypt the encrypted JWT.
Hide the privateKey in a secure location in your server and never tell anyone the privateKey.
I am not a cryptography specialist and hence (I hope) my answer can help somebody who is neither.
There are two possible ways of using cryptography in programming:
Signing / verifying
Encryption / decryption
We use Signing when we want to ensure that data comes from a trusted source.
We use Encryption when we want to protect the data.
Signing / verifying uses asymmetrical algorithms i.e. we sign with one key (private) and the data receiver uses the other (public) key to verify.
A symmetric algorithm uses the same key to encrypt and decrypt data.
The encryption can be done using both symmetric and asymmetric algorithms.
relatively simple article on subject
The above is common knowledge below is my opinion.
When JWT is used for simple client-to-server identification there is no need for signing or asymmetric encryption. JWT can be encrypted with AES which is fast and supersecure. If the server can decrypt it, it means the server is the one who encrypted it.
Summary: non-encrypted JWT is not secure. Symmetric encryption can be used instead of signing in case no third party is involved.
Regarding API call to DocuSign with JWT token,
Can I use my own private key to generate JWT token and share my public key to DocuSign to decode the JWT?
if so, what algorithm to use?
No, you cannot use your own Private Key. You have to use DocuSign provided private Key for JWT authentication.
We have no support in DocuSign JWT for using a customer's public/private keypair. In April, 2019, the product group examined the issue and decided that other feature requests are of higher priority.
If you would like us to look at the idea again, please ask your DocuSign sales person to refer to internal ticket ID-2188.
If I append an HMAC to the auth token then the token cannot be tampered with. I notice many articles also recommend encrypting the token (which contains user id, expiration date and hash). What is the point of encrypting? Isn't that a waste of CPU time? (You have to decrypt the token for each request.) If it is not encrypted, anyone can see the user id and expiration date. What is the harm there?
Using only HMAC is a viable option: http://c2.com/cgi/wiki/wiki?HmacUserAuthentication
But: Note from disadvantages
No privacy. If you need privacy, you need full encryption: use SSL.
HmacUserAuthentication is for security of service provision, not
privacy.
So encryption is a requirement if you use a plain connection. In case of SSL the additional encryption is needed only if you want to prevent the end user from reading the token.
If I have a custom Secure Token Service that specifically lists out allowed audiences and checks if the token is coming from one of of those audiences and also checks the thumbprint and issuer of the X509 certificate, do I need WSFederation?
Since my STS is checking that the the token already came from a specific application and was routed through my ACS, aren't I verifying all of the things I need to? I know that Application A sent a request to the ACS which sent a request to Application B all from the custom STS, so where does Federated Identity fit in this picture?
Edit for clarity:
Sorry I was a bit unclear in the orignal post. I think the confusion came because I used STS instead of security token handler (Way different things, just a typo).
Application A is a custom login service, which displays the login options for the user, google/facebook/yahoo/etc. Logging in through these service gets the token from the ACS and returns it to application B, the Relying Party. This RP has a custom security token handler which accepts the token and validates that it is has an audience URI matching application A. It also validates that the issuer was the ACS and the thumbprint matches the one of the cert used to sign the token via the ACS.
This means that theoretically application B knows, that application A was used to login (as it came from that audienceURI) and that the ACS sent the token (as it was the issuer and the thumbprint matches). What I am asking is if federated identity is necessary for application B? What exactly do you gain by using it, if you've already proved where the token came from?
your question might need some clarification.
First, you might want to explain specifically what you mean by application A and application B, and how your STS fits in this scenario. Applications don't typically issue tokens, only STSes do. In this sense, ACS doesn't connect applications to each other, it connects relying party applications to third party identity providers.
Second, if you're talking about authentication over the web, and you have a custom identity provider STS that's issuing tokens for ACS, then you're probably already using WS-Federation. If however your token acquisition is not browser based, and you're making back-end HTTP calls to ACS, then WS-Federation is not relevant to the scenario.
Third, from the point of view of the STS, the set of allowed audiences is not about token issuers, it refers to entities that will consume tokens issued by that STS. That is, it's the set of subjects that the STS will issue tokens to. This could be applications themselves, or other intermediary STSes along the federation chain. (ACS for example acts as such an intermediary)
Fourth, when you're validating the issuer's certificate on an incoming token, you must do more than just compare the thumbprint. The thumbprint is not part of the token's cryptographic proof. You must validate the token's digital signature in order to verify that the token issuer owns the private key of the certificate.
I hope this clears things up, but if it doesn't answer your question please let me know.
Some time ago we needed a solution for Single Sign On authentication between multiple web services. At least at that time we considered OpenID protocol too complicated and we were not convinced about the Ruby on Rails plugins for it. Therefore we designed a protocol of our own instead of implementing an OpenID provider and OpenID consumers.
I have two questions:
Was it a bad thing not to create
our own OpenID provider and setup
our OpenID consumers accept only it?
Public login or registration are not
allowed and we wanted to keep
authentication simple.
Can you spot a crucial error or a vulnerability in the following design?
If you as a commune can approve this design, I will consider extracting this code into a Ruby on Rails plugin.
Please look at the flowchart and sequence diagram.
Details:
Authentication Provider ("AP"):
Central service which holds all data
about the users.
Only one "AP" exists in this setup.
It could be possible to have multiple "AP"s, but that should not be relevant in this context.
"AP" knows each "S" beforehand.
Authentication Client (Service "S"):
There exists several internal and external web services.
Each service knows "AP" and its public key beforehand.
Actor ("A"):
The end user who authenticates
herself with AP by a username and password
May request directly any URI of "S" or "AP" prior to her login
Connections between "A", "S" and "AP" are secured by HTTPS.
Authentication logic described briefly:
These are a description for the graphical flowchart and sequence diagram which were linked at the top of this post.
1) Auth Provider "AP"
"AP" makes a server-to-server HTTP POST request to "S" to get a nonce.
"AP" generates an authentication token.
Authentication token is an XML entity which includes:
an expiration date (2 minutes from now),
the previously requested nonce (to prevent replay),
identifying name of "S" (token for Service_1 is not good for Service_2),
information about the end user.
Authentication token is encrypted with AES256 and the encryption key and initialization vector are signed by AP's private RSA key.
Resulting strings ("data", "key" and "iv") are first Base64 encoded and then URL encoded to allow them be delivered in the URL query string.
End user "A" is HTTP-redirected to service "S" (HTTPS GET request).
2) Service "S"
Receives authentication token in URL parameters from user agent.
Decrypts authentication token with AP's pre-shared public key.
Accepts one authentication token only once (token includes a nonce which is valid only once).
Checks that identifying name in authentication token corresponds to service's name.
Checks that authentication token is not expired.
Remarks:
It is not a problem if somebody else can also decrypt the authentication token, because it contains no confidential information about the user. However, it is crucial that nobody else than AP is able to generate a valid authentication token. Therefore the RSA key pair is involved.
RSA private key is used only for signing the token, because it cannot encrypt data which is longer than the actual key length. Therefore AES is used for encryption.
Since the authentication token is delivered as an HTTP GET request, it will be stored e.g. in Apache's log file. Using a disposable nonce and an expiration date should minimize the possibility of a replay attack. POST request would need an HTML page with a form which is submitted automatically by Javascript, which is why GET is used.
Service "S" generates a nonce only in a server-to-server API request. Therefore unauthenticated generation requests should not pose a DoS-vulnerability.
You're confusing authentication ("I am who I say I am") and authorization/access control ("I am allowed to access this"). You can just implement OAuth, and then query a server over HTTPS with "is this OAuth identity allowed to access me?". You don't have to worry about replay attacks, since you're using HTTPS.
"Security is hard, so I'll design my own."
Authentication token is encrypted with AES256 and the encryption key and initialization vector are signed by AP's private RSA key.
AES-256 and AES-192 have weak key schedules. But you're not using it for confidentiality; you're using it as some sort of "integrity" check. It doesn't work: Attacker gets a "signed" authentication token. Attacker recovers the key and IV. Attacker encrypts a different authentication token with the same key and IV, and uses the same "signature".
What's wrong with hashing it and signing the hash? Also note that if you're going to use custom signing, you need to be careful about padding (IIRC PKCS-whatever adds at least 11 bytes).
EDIT: And if you're using a cipher where you should be using a hash/MAC, you really shouldn't be designing a security protocol!
Here are a few quick thoughts about question 1:
Designing a working security protocol is very hard, so on general principle I would favor using an existing one.
However, I appreciate that OpenID might not have been very established at the time. Also OpenID is still relatively new and might not have all of its limitations figured out yet.
Still, you'd be using OpenID in a restricted scenario where the big issue of OpenID (involvement of multiple actors) doesn't come into play. You'd only be using the “technical core” of OpenID, which is easier to understand.
Your requirements and the overview of your protocol remind me of Kerberos. I'm also tempted to push towards LDAP + single sign on, but I don't know what concrete solutions exist for that.
A point in favor of your protocol is that you've taken the time to describe it in detail. Just that puts you above than most self-made security protocol designers!
In short I find this protocol to be over engineered in the wrong places and ultimately vulnerable to attack.
So What is the vulnerability?
End user "A" is HTTP-redirected to service "S" (HTTPS GET request).
This is likely to be a violation of OWASP A9. At no point can a user's session ID be passed over an insecure channel such as http. Even if the session id hasn't been authenticated yet, an attacker is patient he can sniff the wire looking for session id's and then periodically check if they have been authenticated and then use them to access your system.
"Complexity is the worst enemy of security."
--Bruce Schneier
Authentication token is encrypted with
AES256 and the encryption key and
initialization vector are signed by
AP's private RSA key.
First of all RSA can be used to encrypt a message, so aes is unnecessary. HTTPS on the other hand is going to be more efficient and proven to be secure. I don't understand why you have to pass an authentication token to the client, when you already have a secure server-to-server communication channel. A server could just say "Hey someone has been redirected to me with this session id, what is his state information?". Its a matter of the weakest link in the chain, and your session id should be strong enough. This would require the session id to be sent as a GET or POST request by the client when the hand off occures which could open the door to Session Fixation. You could check the ip address before and after the handoff, sometimes the ip address of the client can change legitimately but the handoff is going to be a very narrow window in which this can happen, and most importantly it is stops Session Fixation altogether.
In general you should avoid re-inventing the wheal. Especially when it comes to security problems like this which have already been solved. Kerberos is beautiful especially if you need to tie in non-http authentication. Using LDAP for session management is another possibility.
Do you really just sign AES key and then send encrypted token, RSA signature of key and then key-iv in PLAINTEXT?
It's a fail. Attacker can use this key to decrypt a token, change it in any needed way and encrypt back. Your server will never find a difference.
If you want to check token integrity, just make a hash of it and sign this hash. This is what hashes used for. No need to use encryption here.