A fast, light weight and cloud native OAuth 2.0 Server based on microservices architecture built on top of light-4j and light-rest-4j frameworks.
Stack Overflow | Google Group | Gitter Chat | Subreddit | Youtube Channel | Documentation | Contribution Guide |
Light platform follows security first design and we have provided an OAuth 2.0 provider light-oauth2 which is based on light-4j and light-rest-4j frameworks with 7 microservices. Some of the services implement the OAuth 2.0 specifications and others implement some extensions to make OAuth more suitable to protect service to service communication, other styles of services like GraphQL, RPC and Event Driven, Key management and distribution, service registration, token scope calculation and token exchange.
It can support 60000 user login and get authorization code redirect and can generate 700 access tokens per second on my laptop.
It has 7 microservices connected with in-memory data grid and each service can be scaled individually.
OAuth 2.0 is just a specification and a lot of details are in the individual implementation. Our implementation has a lot of extensions and enhancements for additional security and prevent users making mistakes. For example, we have added an additional client type called "trusted" and only this type of client can issue resource owner password credentials grant type.
You can deploy all services or just deploy the services for your use cases. You can deploy token and code service to DMZ and all others internal for maximum security. You can have several token services or deploy token service as sidecar pattern in each node. You can start more instance of key service on the day that your public key certificate for signature verification is changed and shutdown all of the but one the next day. You can take the full advantages of microservices deployment.
The OAuth2 services can be started in a docker-compose for your local development and can be managed by Kubernetes on official test and production environment. It exposes RESTful APIs and can be access from all languages and applications.
Out of the box, it supports Mysql, Postgres and Oracle XE and H2 for unit tests. Other databases can be easily added with configuration change in service.yml.
With distributed security verification, JWT signature public key certificates must but distributed to all resource servers. The traditional push approach is not working with microservices architecture and pull approach is adopted. There is a key service with endpoint to retrieve public key certificate from microservices during runtime based on the key_id from JWT header.
Each service in a microservices application needs a subject token which identifies the original caller (the person who logged in the original client) and an access token which identifies the immediate caller (might be another microservices). Both tokens will be verified with scopes to the API endpoint level. Additional claims in these tokens will be used for fine-grained authorization which happens within the business context.
Even with two tokens, we can only verify who is the original calller and which client is the immediate caller. For some highly protected service like payment or fund transfer, we need to ensure that the call is routed through some known services. light-oauth2 token service support token exchange and chaining so that a service can verify the entire call tree to authorize if the call is authorized or not.
light-oauth2 has a service registration to allow all service to be registered with service id and all endpoints as well as scopes for the endpoint. During client registration, you can link a client to services/endpoints and the scope of the client can be calculated and updated in client table. This avoids developers to pass in scopes when getting access token as there might be hundreds of them for a client that accesses dozens of microservices.
A database audit handler has been wired into all light-oauth2 services to log each activity across services with sensitive info masked. In the future we will put these logs into AI stream processing to identify abnormal behaviors just like normal service log processing.
light-java to build API
light-oauth2 to control API access
light-portal to manage clients and APIs
This introduction document contains all the basic concept of OAuth 2.0 specification and how it work in general.
The easiest way to start using light-oauth2 in your development environment is through docker-compose in light-docker repository. Please refer to getting started for more information.
There are some key decision points that are documented in architecture section.
The detailed service document help users to understand how each individual service works and the specification for each services. It also contains information on which scenarios will trigger what kind of errors.
There are tutorials for each service that shows how to use the most common use cases with examples.
There are vast amount of information about OAuth 2.0 specifications and implementations. Here are some important references that can help you to understand OAuth 2.0 Authorization.