Skip to content

Latest commit

 

History

History
83 lines (72 loc) · 3.52 KB

README.md

File metadata and controls

83 lines (72 loc) · 3.52 KB
Raw

CQRS Cache

Codacy Badge CircleCI codecov

Introduction

Build REST API with CQRS (Command Query Responsibility Segregation) architecture to store key-value in local memory cache. This is a sample project using CQRS design.

Requirements:

  • Access data as normal cache (set and get methods)
  • Throttle request per ip-address or username
  • Persistence data

Normally we're easy to catch CQS (Command Query Seperation) than CQRS. CQS puts commands and queries in different methods within a type otherwhile CQRS puts commands and queries on different objects. In this project, I separate data into 2 objects:

  • Raw Data: using to update cache
  • Aggregate Data: using to get the latest cache state (Ex: number of requests by ip-address)

API

  • POST /cache/add
  • POST /cache/get
  • POST /cache/remove
  • GET /cache/peek
  • POST /cache/take

Get how many request send to cache in interval time by ip-address. Interval time is configured by rate-schedule in application.conf

  • GET /cache/rate?ipAddress=
  • GET /cache/rate-report

Body format:

{  
    "key" : string,  
    "value" : string  
}

For example:

{  
    "key":"01234567-9abc-def0-1124-56789abc1004",  
    "value":"1234"  
}

Technologies

  • Play framework 2.6
  • Akka 2.5.6 (Persistent Actor)

Design

Architecture

  • Using CQRS
           add/remove/take/peek  ┌────────────────┐       ┌──────────────────┐
                           ┌───▶ │ CommandService │ ─────▶│ RawInMemoryActor │  
                           │     └────────────────┘       └──────────────────┘ 
                           │              |
     ┌─────────────────┐   │              |        forward
──▶  │ CacheController │──▶│              └--------------------------┐                 
     └─────────────────┘   │                                         ▼
                           │     ┌──────────────┐         ┌────────────────────────┐
                           └───▶ │ QueryService │ ──────▶ │ AggregateInMemoryActor │
            rate/rate-report     └──────────────┘         └────────────────────────┘

Data structure

  • The data structure in the project based on MRU (Most Recently Used) cache and used the LinkedMap to implement the key-value memory cache. LinkedMap keeps track of the order in which each element is added, so the complexity of Add/Remove/Peek/Take is O(c) (constant time).
  • Using PersistentActor to store persistent data in Database.

Running

  • Start docker
docker-compose up -d
  • Start application
sbt run