To start service on your local machine:
./mvnw spring-boot:run
To start all tests:
./mvnw clean test
Assumptions made
I've created the CompanyRepository interface that is an abstraction for data storage.
For a case of simplicity of an exercise I've implemented in-memory storage CompanyRepositoryInMemory
In the production ready application the CompanyRepository interface should be implement as
database backed storage.
CURL
To create new company:
curl -X POST -H "Content-Type: application/json" -d @exampleCompany.json https://company-micro-service.herokuapp.com/company
To get company:
curl -i -H "Accept: application/json" http://localhost:8080/company/1
To get all companies:
curl -i -H "Accept: application/json" http://localhost:8080/companies
To add beneficial owner for the company:
curl -X POST -H "Content-Type: application/json" -d @beneficialOwner.json http://localhost:8080/company/1/beneficial-owner
Application is deployed to heroku:
host:
https://company-micro-service.herokuapp.com/
So you all curl queries could be executed against that host, for example:
curl -X POST -H "Content-Type: application/json" -d @exampleCompany.json https://company-micro-service.herokuapp.com/company
and
curl -i -H "Accept: application/json" https://company-micro-service.herokuapp.com/company/100
Authentication:
For Authentication I would use OAuth2.
Client will use client_secret and client_id to get token.
Backend will provide client with the token.
Next, token will be used and passed as the Authentication header with every https request.
The Authentication token will have expiration time.
When expiration time will elapsed client will request for renewal of token and
the process will repeat.
Every request should be done via HTTPs
How can you make the service redundant? What considerations should you do?
- We should deploy our application to multiple nodes(hosts). Then when one host will have problems we can redirect whole traffic to other hosts. Between clients and our backend services should be a load balancer (e.g. varnish) that is routing traffic to only healthy instances.
- If we are using i.e. AWS for our infrastructure we can deploy some instances in one zone(dc) i.e. europe and other in other dc to reduce probability of failure if one dc is down