Spring Cloud Microservices - Movie Recommendation System using Google's Protocol Buffers
The application includes the following microservices.
- Eureka Discovery Service
- Config Service
- User Service
- Movie Service
- Recommendation Service
- Recommendation Client Service
The application includes .proto files for entities that are sent between services
- User
- Movie
- ReccomendationData
Proto entities are created using protoc compiler which is included in project.
Microservices communicate over network using Protocol Buffers, and for human readability recommendation-client service convert protobuff to the JSON at the last endpoint.
This reference application uses the following mixture of databases.
- MySQL - RDBMS
- Neo4j - GraphDB
- MongoDB - Document Store
Use Postman to make request. Collection of all request are exported and can be found in root of this repository.
Configuration service MUST be the first service to be started, only after that, other services can be started. To run configuration service, open terminal in root directory and execute:
mvn spring-boot:run
It will run on localhost:8888. This service connects to [this] (https://github.com/mdeket/spring-cloud-example-config-repo) github repository and loads property files for other services.
Discovery service is the one of key tenents of a microservice based architecture. Eureka is the Netflix Service Discovery Server and Client. To run this service, do the same as for Config Service, open terminal, position it to discovery service root directory and execute:
mvn spring-boot:run
Eureka starts on localhost:8761.
Contains all users for this system. They are all stored in UserSOA database. Before starting user service, you must have MySQL installed and running on localhost:3306 with UserSOA database created. All other db properties are stored here To run user service, do the same as for the last two services.
By starting the app, spring runs data.sql script with 50 users, with names and emails. Script can be found in resource folder. App will run on localhost:8001.
Is in charge of movies. Before starting this service, just be sure to have installed and a running mongodb instance on default port. To populate database, run Movie Service - Dummy Data request from postman collection you imported. To start movie service, do the same as for last three services, just position yourself inside of movie service project 😁 .
Is in charge of recommendation logic. Its database contains basic information about users(id) and movies(id), and as well as information about which user liked which movie and who is user following. Neo4j must be installed in order to run this service. After running this service (localhost:9000) and execution of Dummy Data request from postman, you can see all the nodes and paths in Neo4j Graph Visualization that runs on localhost:7474.
If user for which recommendation is requested is not following anybody, he/she will get top 5 movies based on likes.
Represents an API Gateaway for all services. It receives http requests from user's browser and then it passes request to other services. It uses ribbon for client side load balancing, so for example two services of recommendation service can be up, and ribbon will take take which instance to use on each request. To be able to do that, you have to change configuration of recommendation service so you could be able to run instance twice on different ports. Recommendation client uses circiut breaker (hystrix). So if recommendation service is down or movie service is down, it has 5 default movies that will be used as recommendation.
Protocol buffers are a language-neutral, platform-neutral extensible mechanism for serializing structured data.
Project includes stress_test.py script for benchmark test for 50 request for proto endpoint and 50 request for non proto endpoint. Results may vary from 4-5 sec to 1 sec for 50 requests.
My results :
python stress_test.py
Proto : 1.25499987602
Json : 1.22300004959