Update/change roles claim (or any other claim) in JWT - authentication

I'm storing user roles inside a JWT (to restrict API endpoints). The roles can be changed by an administrator.
If a role is changed. How am I supposed to reflect this inside all tokens? I've thought about a couple of solutions:
If I'd use refresh tokens, the user would have to wait until the expiration date of the access token is expired.
I could keep a record of changed user IDs and check every request, and then return a new token if the user has been changed.
Is there a standard way to do this?

Refresh tokens don't seem to be the solution if you care about the changes you make being instant, you probably don't want an user to access moderation tools for some time if you revoke his permissions.
What you could do is keep a version number in the jwt token relative to the user, much like how mongoose does it with it's versionKey. By doing this, you would be able to check this version against the one in the database for a given user. Each time you change the roles of this user, you would increment this version, if the version of the jwt doesn't match, just recreate a new one with the correct roles and version and send it back to the user.
I don't believe there is a proper standard for this, as jwt is immutable by design, you'll have to change it entirely if you need to "update" it.

The JWT tokens are immutable so you can't change/update claims on an existing token - thus you have to issue a new JWT token.
That leads to the biggest problem with JWT - token revocation. There are no good solutions. What you can do is
Keep JWT expiration date short (and optionally use refresh tokens)
Use a blacklist to keep a list of revoked tokens (of course losing the 'stateless' part this way)
change the secret key (keep in mind that this revokes ALL valid tokens of all users)
The best solution depends on the concrete case.

To address that scenario, you can keep the token lifetime short, once the token expires you can renew the token silently in case of Implicit grant or use refresh token mechanism to issue a new token from the authorization server. Once the role has changed, it might not reflect in the token instantly, but it will be available once the token is renewed.

#Janar said three solutions to handle different scenarios, but they still exist corresponding drawbacks.
I also have a idea to revoke your token:
set iat to JWT payload. (You should know what is the iat)
revoke tokens of a user:
find the user id
let auth server rejects tokens which are before current time(judged by iat) and belong to this user(judged by userId)

A solution to this would be to keep authentication and authorization separate when using JWT tokens. Use the token for authentication(claiming an ID) and use that ID to check the authorization in the database. This can increase the latency of each request considering that you now have to check the authorization for each request. However, this can be mitigated by using an in-memory data structure store(such as Redis) as a cache. On every permission change, delete the cache and make another call for the permissions.

Related

Why should I store a JWT refresh token server-side?

I've read several articles (such as this and this, and this SO answer) that suggest storing refresh tokens server-side in a database, or in-memory (or an in-memory store, such as Redis). From what I understand this is so they can be revoked.
Is there a good reason for storing all tokens as these articles suggest, rather than just storing blacklisted tokens on logout? If I understand the reasons for storing tokens, surely I could achieve the same effect by storing a token id in Redis (with a TTL as long as the expiry, so that the table doesn't grow unweildy).
Are there downsides to this approach, and if so, what are they (or conversely, what are the upsides of storing all tokens, vs just a list of revoked tokens?)
To elaborate, and why I think it should be fine to have a revocation list, here's the process I am imagining:
Issue tokens out
Once revoked (say, on logout), add an entry to a blacklist for a unique id (say, public_user_id if there are many tokens for different devices) with the revocation time, and add a TTL until the token's expiry
When a token is provided for auth, when there is an existing blacklist entry:
if a valid token is used, it'll be before it's expiry and have a creation time after the blacklist entry creation time for the unique id
if an invalid token is used, it'll either be expired or within the blacklist with the unique identifier before the entry creation
Am I missing something critical in that flow that would require a list of all tokens instead?
The advantage of having a list of all issued tokens is that you can have a full view of who has been already authenticated and has currently access to the system. You can then choose to revoke some tokens from this list based on any criteria (e.g. the age of the token, the roles associated with the user of the token, the IP address ranges).
If you only have a list of revoked tokens, it would be impossible to choose at runtime, an arbitrary criteria, to revoke a subset of the valid tokens. Stated otherwise, if you don't have a list of all issued tokens, the revocation criteria cannot be enforced globally at once, but only when a token is presented to a resource server.
The only deciding factor I can think of, that will require a list of all refresh tokens is the following:
Do you / will you, at any point, need to have a functionality where you can dynamically revoke valid refresh tokens, based on some arbitrary, regulatory, legal, integrity, security etc. criteria?
If so, the least you will need is a list of all issued tokens, plus any metadata required to implement the criteria logic.
Example: "Due to regulation, I need to ban all EU users" equates to delete from refresh_tokens were user_ip in <... eu logic ...>
I would aim for a direction of technical simplicity here, since the Authorization Server (AS) should do the hard work for you. Here are some end to end notes which explain some tricky aspects and suggest a simple direction.
1. TOKEN ISSUING
The user authenticates (and optionally consents) resulting in a token 'grant'. The AS should then store refresh tokens for you, in a database table that might be named 'delegations'. Typically the stored 'token' will be a hash rather than the real value, and will be linked to the application (client_id) and user (subject). Tokens issued might have these lifetimes:
Refresh token: 4 hours
Access token: 30 minutes
2. TOKEN REFRESH
OAuth clients such as mobile apps will silently renew access tokens during the lifetime of the grant or 'user session'. This involves sending a refresh token to the AS and getting back a new access token. For this to work the AS needs to store a hash of the refresh token in order to be able to validate the input.
3. DEFAULT REMOVAL BEHAVIOUR
When a user logs out, tokens are cleared from the client app, so they are gone. Newer OAuth 2.1 recommendations are to use rotating refresh tokens, where each access token refresh also renews the refresh token and invalidates the previous one. In our example this now means that the lifetime of a stolen refresh token is likely to be reduced - perhaps to only 30 minutes.
4. MANUAL REVOCATION BY ADMINISTRATOR
If for some reason you want to explicitly deny access to a particular user and application, an administrator could use the AS database and issue a command like this, though the Admin UI may provide more visual options.
delete from delegations where client_id=[value] and subject=[value]
Whether manual revocation is likely to be manageable at a people level is questionable but it is a good capability to have, eg in security reviews.
5. REVOCATION OF REFRESH TOKENS ON LOGOUT
Of course the client can revoke its own refresh tokens on logout if required, before clearing tokens. This should also ensure that any access tokens for the same grant are rejected by the Authorization Server.
6. ACCESS TOKEN VALIDITY
Access tokens are most commonly JWTs. Revocation or logout may occur when the JWT still has 25 minutes to live. If an attacker has somehow intercepted an access token (which shouldn't usually be possible) they can continue to use it against your APIs - during this period the AS will never see the access token.
7. API GATEWAY SETUPS
In a more sophisticated setup, opaque access tokens are issued to internet clients, then sent to an API gateway, which introspects them, as in the Phantom Token Pattern. The gateway also maintains a cache of the access token result.
At the time of revocation the AS can raise a Custom Event to inform the API Gateway, which can then clear any cached access tokens for the user. This should ensure that the very next request with an access token for the revoked (or logged out) user is rejected.
SUMMARY
Unless you are dealing with a very high security domain I would aim for simplicity and follow these two principles:
Leave token storage and revocation to the Authorization Server
Keep tokens short lived so that revocation becomes less of an issue

Manually expire JWT token from server - WEB API 2

I am working on a api server which revives requests from a mobile app. I am using JWT with ASP.Net MVC Web API 2. In this Admin gives access of various departments to mobile app users. I set these DeptIds in Claims at the time of Login. On every authorised request from app, in a custom action filter attribute I read claims to match deptId in request URL with claims. This all scenario works fine.
Now my problem is, when Admin revokes access of any particular dept from app user, how should I expire the access_token of that user so that, on Login request call, I can set new Claims. Otherwise, as Admin removes access from server but the deptId still exists in user's Claims so user still have access to that department.
One way is on every request, check in database for access but that increases overhead of server also increases response time. So I don't want to go this way.
I didn't find anything on web how to expire token in JWT. Can any one help on this?
A JWT token happens to be a kind of token that allows it to be self-contained i.e. one can validate it without consulting an external entity.
That also means that there's no external entity that will tell you that the token was revoked or expired. Expiration can only happen by some property in the JWT itself; the exp claim was standardized for that purpose: it will tell the recipient the time after which the information in it must no longer consider to be valid
Authentication and Authorization are different things.
Use JWT for Authentication but not for Authorization. I mean that using JWT you can know who the user are but don't put information about what user can do into the JWT. Check permissions for the user on the server side basing on just user's identity that you got from JWT. You may also put into JWT some information that additionally limits access rights (i.e. blacklisting; for example, social network site creates an access token for a game to access my identity and friends list but not my posts) but don't put there information that directly provides access to some features (i.e. whitelisting). In that way you can easily remove access to some features on your sever-side despite that fact that the user is already logged in with JWT.
I understand you are interested in revoking or invalidating tokens instead of expiring them.
Revoking or invalidating tokens
Unfortunately​ it's not possible to achieve it
without keeping the track of the tokens somewhere:
Issue a token and add it to a whitelist to keep the track of it.
When validating a token, check the whitelist and refuse the request if it is not whitelisted.
To revoke or invalidate a token, remove it from the whitelist.
This security schema require some trade-offs. Deal with it.
Performance considerations
Regarding your performance concerns: Bear in mind that premature optimization is the root of all evil. You shouldn't optimize until you have a performance problem and you have proven that the performance problem comes from the way you store your tokens.
You could start storing the tokens in the database and then consider a cache in memory, for example. But always be careful when fixing a problem that you currently don't have.
Token identifier
The jti claim should be used to store the token identifier on the token. When validating the token, ensure that it's valid by checking the value of the jti claim against the token identifiers you have on server side.
For the token identifier you could use a UUID.

Security and reliability concerns about JSON Web Tokens (JWT)

I'm creating an API along with a SPA for a personal project and I'm hesitating between the following solutions to authenticate users (note: over HTTPS):
HTTP Basic Authentication (send username/password wich each request)
Token based authentication (store SHA1-ed user tokens in the database)
JSON Web Token (JWT) authentication
I don't even consider OAuth cause it seems like a real pain plus I don't need to authenticate with other apps, I'm only concerned about authenticating users.
From what I've read, JWT seems to be a growing standard. It basically holds the caller's data so everytime he makes an API request you encrypt(base64(header) + "." + base64(payload)) with your secret and you compare it with the signature provided in the last part of the token itself. It avoids having to perform DB transactions.
The problem is that if I use JWT 1) I have no possibility to manually revoke specific tokens, and most of all 2) if I change a user's permissions, the previously granted JWT will still have the old data with his old permissions which could grant/restrict him continuous access to some data as long as he doesn't get a new token with his new permissions, which is really problematic and I'm surprised I haven't seen anyone mentionning this problem yet. Moreover, 3) JWT claims to allow the server to validate access without having access to DB but I can't imagine any API request that doesn't involve the database somehow, if only to return data the user asked for. So this argument doesn't make any sense to me.
To me, my best option right now is option 2. Website will have restricted and small traffic so storing tokens in the Database seems like a small and worthwhile trade-off and allow me to do anything I want with these tokens, including managing their lifecycle and permissions. It also avoids exposing the users' credentials like in option 1, in case they use the same ones for other online services.
I just want to know if my concerns about JWT are right or if I misunderstood its functioning? Also, even if I've already read a lot about these different options, feel free to link anything that could enlight me and help me make a better choice. Thanks.
You are right and invalidating tokens before expiration time is a common JWT problem. There are several reason to consider: account deleted/blocked/suspended, password changed, permissions changed, user logged out by admin.
With JWT, you can also set a token blacklist to store tokens that were between logout & expiry time, mark expired and check it in every request. You can include only the ID (jti claim of JWT) or use the last login date and the iat claim (issued at)
Other technique to invalidate tokens when user changes their password/permissions is signing the token with a hash of those fields. If the field value changes, any previous tokens automatically fail to verify.
See https://stackoverflow.com/a/37520125/6371459
Finally, be aware that the token is signed with server private key (not encrypted)
sign(base64(header) + "." + base64(payload))

Is a Refresh Token really necessary when using JWT token authentication?

I'm referencing another SO post that discusses using refresh tokens with JWT.
JWT (JSON Web Token) automatic prolongation of expiration
I have an application with a very common architecture where my clients (web and mobile) talk to a REST API which then talks to a service layer and data layer.
I understand JWT token authentication, but I am a little confused at how I should use refresh tokens.
I want my JWT authentication to have the following properties:
JWT Token has an expiration of 2 hours.
The token is refreshed every hour by the client.
If the user token is not refreshed (user is inactive and the app is not open) and expires, they will need to log in whenever they want to resume.
I see a lot of people claiming to make this a better experience using the concept of a refresh token, however, I don't see the benefit of this. It seems like an added complexity having to manage it.
My questions are the following:
If I WERE to use a refresh token, wouldn't it still be beneficial to have a long term expiration for good practice on that token as well?
If I WERE to use a refresh token, would that token be persisted with the userId and/or JWT token?
When I update my token every 1 hour, how does this work? Will I want to create an endpoint that takes in my JWT token or my refresh token? Will this update the expiration date of my original JWT token, or create a new token?
Is there the need for a refresh token given these details? It seems that If the user is just using a JWT token to grab a new token (per the link above) then the refresh token is obsolete.
Let me come to your questions a little later down the line and start by actually discussing the whole purpose of a refresh token.
So the situation is:
The user opens the app and provides his login credentials. Now, most probably the app is interacting with a REST backend service. REST is stateless, there isn't a way to authorize access to the APIs. Hence, so far in the discussion, there is no way to check if an authorized user is accessing the APIs or is just some random requests coming through.
Now to be able to solve this problem, we need a way to know that the requests are coming from an authorized user. So, what we did was to introduce something called an access token. So now once the user is authenticated successfully, he is issued an access token. This token is supposed to be a long and highly random token (to ensure that it can not be guessed). This is where the JWT comes into the picture. Now you may/may not want to store any user-specific details in a JWT token. Ideally, you would want to just store very simple, extremely non-sensitive details in the JWT. The manipulation of the JWT hash to retrieve other user's details (IDOR etc.) is taken care of by JWT (the library being used) itself.
So, for now, our problem with authorized access is solved.
Now we talk of an attack scenario. Let's say using all of the above user Alice, using the app, has the authorized access token and now her app can make requests to all the APIs and retrieve the data as per her authorization.
Assume that SOMEHOW Alice loses the Access Token or put another way, an adversary, Bob, gets access to Alice's access token. Now Bob, despite being unauthorized, can make requests to all the APIs that Alice was authorized to.
SOMETHING WE IDEALLY DON'T WANT.
Now the solution to this problem is :
Either detect that there is something of this sort happening.
Reduce the attack window itself.
Using just the access token alone, it is hard to achieve condition 1 above, because be it Alice or Bob, it's the same authorized token being used and hence requests form the two users are not distinguishable.
So we try achieving 2 above and hence we add an expiration to the validity of the access token, say the access token is valid for 't' (short-lived) time.
How does it help? Well, even if Bob has the access token, he can use it only while it is valid. As soon as it expires, he will have to retrieve it again. Now, of course, you could say that he can get it the same way he got it the first time. But then again there's nothing like 100% security!
The above approach still has a problem and in some cases an unacceptable one. When the access token expires, it would require the user to enter his login credentials and obtain an authorized access token again, which at least in case of mobile apps, is a bad (not acceptable) user experience.
Solution: This is where the refresh token comes in. It is again a random unpredictable token that is also issued to the app along with the access token in the first place. This refresh token is a very long-lived special token, which makes sure that as soon as the access token expires, it requests the server for a new access token, thus removing the need for the user to re-enter his login credentials to retrieve a new authorized access token, once an existing one has expired.
Now you may ask, Bob can have access to the refresh token as well, similar to the way he compromised the access token. YES. He can. However, now it becomes easy to identify such an incidence, which was not possible in the case of an access token alone, and take the necessary action to reduce the damage done.
How?
For every authenticated user (in case of a mobile app, generally), a one to one mapped refresh token and access token pair is issued to the app. So at any given point in time, for a single authenticated user, there will be only one access token corresponding to a refresh token. Now assume that if Bob has compromised the refresh token, he would be using it to generate an access token (because access token is the only thing which is authorized to access resources through the APIs). As soon as Bob (attacker) requests with the newly generated access token because Alice's (genuine user) access token is still valid, the server would see this as an anomaly, because for a single refresh token there can be only one authorized access token at a time. Identifying the anomaly, the server would destroy the refresh token in question and along with it all, it's associated access tokens will also get invalidated. Thus preventing any further access, genuine or malicious, to any authorization requiring resources.
The user, Alice, would be required to once again authenticate with her credentials and fetch a valid pair of a refresh and access tokens.
Of course, you could still argue that Bob could once again get access to both refresh and access tokens and repeat the entire story above, potentially leading to a DoS on Alice, the actual genuine customer, but then again there is nothing like 100% security.
Also as a good practice, the refresh token should have an expiry, although a pretty long one.
I believe for this scenario you could work with the access token alone, making
life easier for your clients but keeping the security benefits of a refresh token.
This is how it would work:
When your user logs in with credentials (username/password) you return a
short-lived JWT. You also create a db record where you store:
JWT id
user id
IP address
user agent
a valid flag (defaults to TRUE)
createdAt
updatedAt
Your client submits the JWT in every request. As long as the JWT hasn't expired,
it has access to the resources. If the JWT expired, you refresh it
behind the scenes and return both the resource and an additional X-JWT header
with the new JWT.
When the client receives a response with an X-JWT header, it discards the
old JWT and uses the new one for future requests.
How refreshing the JWT works on the server
Look for the matching db record using the JWT id.
Check if the valid flag is still true, otherwise reject.
Optionally, you can compare the request IP address and user agent against
the stored IP address and user agent, and decide to reject if something looks
fishy.
Optionally, you can check the db record's createdAt or updatedAt fields, and
decide not to refresh if too much time has passed.
Update the updatedAt field in the db record.
Return the new JWT (which is basically a copy of the expired JWT, but with an extended expiration time).
This design would also give you the option to revoke all tokens for a user (for
example, if the user loses his phone or updates his password).
Benefits:
Your client never has to check expiration times or make refresh token
requests, all it does is check for an X-JWT header on responses.
You can add custom refresh logic based on IP address, user agent, max-token
age, or a combination of those.
You can revoke some or all tokens for a user.
If I WERE to use a refresh token, wouldn't it still be beneficial to have a long term expiration for good practice on that token as well?
Refresh Tokens are long-lived, Access Tokens are short-lived.
If I WERE to use a refresh token, would that token be persisted with the userId and/or JWT token?
It would be persisted as a separate token on the client, alongside JWT but not inside JWT. UserID/UID can be stored inside the JWT token itself.
When I update my token every 1 hour, how does this work? Will I want to create an endpoint that takes in my JWT token or my refresh token? Will this update the expiration date of my original JWT token, or create a new token?
Yes, you need a separate service that issues and refreshes token. It won't update the expiration of the existing JWT Token. A token is simply JSON field-value pairs that are base64 encoded. So changing the data, changes the output. The token also has the issue date, which will at the very least change on every fresh issue (refresh). So every token will be unique and new. The old tokens will auto-expire, hence you need expiration on all Access Tokens, otherwise they will linger around forever.
The other answer here states that old tokens get destroyed when you issue a new token. That's simply not the case. Tokens cannot be destroyed. In fact, you can harvest hundreds of tokens by constantly contacting the auth server and asking for new fresh tokens using your Refresh Token. Each of those Access Tokens will be valid till their expiry. So expiry is imperative, and it should be short.
Is there really the need for a refresh token given these details? It seems that If the user is just using a JWT token to grab a new token (per the link above) then the refresh token is obsolete.
JWT tokens have client claims. For example is_manager:true claim on a JWT token might allow access to manager-level features. Now if you decide to demote the user from manager to contractor, that won't take effect immediately. The user may still be using the old token. Finally when that expires, he hits the auth server to refresh his token. The auth server issues a new token without the managerial claim and the user won't be able to access managerial features any more. This creates a window during which the user's claims are not in sync with the server. This again explains why Access Tokens should be short-lived so sync'ing can happen often.
Essentially you are updating the authorization checks every 15 minutes, instead of checking them on every single request (which is how typical session-based auth works). If you want real-time permissions instead of every-15-minute refreshes, then JWT may not be a good fit.

what's the point of refresh token?

i have to confess i've had this question for a very long time, never really understand.
say auth token is like a key to a safe, when it expires it's not usable anymore. now we're given a magic refresh token, which can be used to get another usable key, and another... until the magic key expires. so why not just set the expiration of the auth token as the same as refresh token? why bother at all?
what's the valid reason for it, maybe a historical one? really want to know. thanks
I was reading an article the other day by Taiseer Joudeh and I find it very useful he said:
In my own opinion there are three main benefits to use refresh tokens which they are:
Updating access token content: as you know the access tokens are self contained tokens, they contain all the claims (Information) about the authenticated user once they are generated, now if we issue a long lived token (1 month for example) for a user named “Alex” and enrolled him in role “Users” then this information get contained on the token which the Authorization server generated. If you decided later on (2 days after he obtained the token) to add him to the “Admin” role then there is no way to update this information contained in the token generated, you need to ask him to re-authenticate him self again so the Authorization server add this information to this newly generated access token, and this not feasible on most of the cases. You might not be able to reach users who obtained long lived access tokens. So to overcome this issue we need to issue short lived access tokens (30 minutes for example) and use the refresh token to obtain new access token, once you obtain the new access token, the Authorization Server will be able to add new claim for user “Alex” which assigns him to “Admin” role once the new access token being generated
Revoking access from authenticated users: Once the user obtains long lived access token he’ll be able to access the server resources as long as his access token is not expired, there is no standard way to revoke access tokens unless the Authorization Server implements custom logic which forces you to store generated access token in database and do database checks with each request. But with refresh tokens, a system admin can revoke access by simply deleting the refresh token identifier from the database so once the system requests new access token using the deleted refresh token, the Authorization Server will reject this request because the refresh token is no longer available (we’ll come into this with more details).
No need to store or ask for username and password: Using refresh tokens allows you to ask the user for his username and password only one time once he authenticates for the first time, then Authorization Server can issue very long lived refresh token (1 year for example) and the user will stay logged in all this period unless system admin tries to revoke the refresh token. You can think of this as a way to do offline access to server resources, this can be useful if you are building an API which will be consumed by front end application where it is not feasible to keep asking for username/password frequently.
I would like to add to this another perspective.
Stateless authentication without hitting the DB on each request
Let's suppose you want to create a stateless (no session) security mechanism that can do authentication of millions of users, without having to make a database call to do the authentication. With all the traffic your app is getting, saving a DB call on each request is worth a lot! And it needs to be stateless so it can be easily clustered and scaled up to hundreds or even thousands of servers.
With old-fashioned sessions, the user logs in, at which point we read their user info from the database. To avoid having to read it again and again we store it in a session (usually in memory or some clustered cache). We send the session ID to the client in a cookie, which is attached to all subsequent requests. On subsequent requests, we use the session ID to lookup the session, that in turn contains the user info.
Put the user info directly in the access token
But we don't want sessions. So instead of storing the user info in the session, let's just put it in an access token. We sign the token so no one can tamper with it and presto. We can authenticate requests without a session and without having to look up the user info from the DB for each request.
No session ... no way to ban users?
But not having a session has a big downside. What if this user is banned for example? In the old scenario we just remove his session. He then has to log in again, which he won't be able to do. Ban completed. But in the new scenario there is no session. So how can we ban him? We would have to ask him (very politely) to remove his access token. Check each incoming request against a ban list? Yes, would work, but now we again have to make that DB call we don't want.
Compromise with short-lived tokens
If we think it's acceptable that a user might still be able to use his account for, say, 10 minutes after being banned, we can create a situation that is a compromise between checking the DB every request and only on login. And that's where refresh tokens come in. They allow us to use a stateless mechanism with short-lived access tokens. We can't revoke these tokens as no database check is done for them. We only check their expiry date against the current time. But once they expire, the user will need to provide the refresh token to get a new access token. At this point we do check the DB and see that the user has been banned. So we deny the request for an access token and the ban takes effect.
The referenced answer (via #Anders) is helpful, It states:
In case of compromise, the time window it's valid for is limited, but
the tokens are used over SSL, so unlikely to be compromised.
I think the important part is that access tokens will often get logged (especially when used as a query parameter, which is helpful for JSONP), so it's best for them to be short-lived.
There are a few additional reasons, with large-scale implementations of OAuth 2.0 by service providers:
API servers can securely validate access tokens without DB lookups or RPC calls if it's okay to not worry about revocation. This can have strong performance benefits and lessen complexity for the API servers. Best if you're okay with a token revocation taking 30m-60m (or whatever the length of the access token is). Of course, the API servers could also keep an in-memory list of tokens revoked in the last hour too.
Since tokens can have multiple scopes with access to multiple different API services, having short-lived access tokens prevents a developer of API service for getting a lifelong access to a user's data on API service B. Compartmentalization is good for security.
Shortes possible answer:
Refresh tokens allow for scoped / different decay times of tokens. Actual resource tokens are short lived, while the refresh token can remain valid for years (mobile apps). This comes with better security (resource tokens don't have to be protected) and performance (only the refresh token API has to check validity against DB).
The following is an addition to the benefits of refresh tokens that are already mentioned.
Safety First!
Access tokens are short-lived. If someone steals an access token, he will have access to resources only until access token expires.
"...But what if a refresh token is stolen?"
If an attacker steals the refresh token, he can obtain an access token. For this reason, it it recommended that a new refresh token is issued each time a new access token is obtained. If the same refresh token is used twice, it probably means that the refresh token has been stolen.
When the refresh token changes after each use, if the authorization
server ever detects a refresh token was used twice, it means it has
likely been copied and is being used by an attacker, and the
authorization server can revoke all access tokens and refresh tokens
associated with it immediately.
https://www.oauth.com/oauth2-servers/making-authenticated-requests/refreshing-an-access-token/
Of course, this is just another layer of security. The attacker can still have time to obtain access tokens, until the refresh token is used a second time (either by the attacker or the real user).
Always keep in mind that the refresh token must be stored as securely as possible.