improve login security by learning and checking additional features like typing speed
This distributed application is responsible for:
- collecting training data:
collector-service - anomaly detection using the multivariate gaussian distribution:
learning-service - validating password basing on features like typing speed:
validator-service - authenticating user:
login-service
- distributed, based on micro-services application written in Spring Boot & Cloud - goal: improve Spring's skills and writing multi-services applications
- no external machine learning libraries, all algorithms written using only basic math formulas (
+,-,*,/,^) - goal: know how anomaly detection algorithms work - use any Continuous Integration and Delivery systems together with tests - goal: know how to configure CI/CD systems and keep code quality discipline
- deploy as docker's containers on any cheap VPS and make the application available for everyone - goal: know how docker deployment works from production point of view
http://ml-login.ceszke.com (currently offline)
- Spring Boot
- Spring Cloud (Config, Ribbon, Feign)
- Spring Security
- Spring Data JPA (Hibernate, H2)
- JUnit, Spring MVC Test
- Thymeleaf
- Twitter Bootstrap
- Maven
- Travis CI
- Docker
- prepare GUI:
- login page ✔
- integrate with Spring Security ✔
- screenshots ✔
collector-serviceapi ✔- basic use case... ✔
- implement algorithm for choosing
εinlearning-service✔ - implement algorithm for model's evaluation in
learning-service✔ - validation of the input (request) data (in controllers)
- introduce exception handling ✔
- commons - split, refactor, clean up
- introduce
-clientmodules - add hystrix, zuul, eureka
- add config clients/server ✔
- extract configuration ✔
- extract paths
- clean up maven poms (introduce hierarchy)
- refactor: service names ✔
- introduce wiremock and integrational tests (multi services)
- write more tests ✔
- replace Thymeleaf by any modern frontend framework (angular/react/vue?)
- prepare live demo ✔