Microservice Basics

(or: what I have learned in the last few weeks)

This is a small collection of software pieces with the intention to show the use of different architecture elements in cloud-based solutions with implementation in different programming languages: C, Java, Python.

Overview

A quite simple web site shows on port 8080 four active elements:

• a counter field that is increased every second

• a Reset Button that sets the counter to 0

• a latest counter field that shows the counter value at the time of reset (+ timestamp + possible errors)

• an average field that shows the average of the last 10 reset counters.

The web site gets it's data from a microservice API Gateway which gets it's data from three different services:

• ServiceJava for the counter value

• ServiceC for the lastcounter value

• ServicePython for the Average of the latest 10 counters (namend summary).

Elements

The elements I intend to show in this small examples are:

Element	Description	Link
Docker	Some kind of virtual machines ("Containers") to keep microservices in a closed environment	https://docs.docker.com/reference/
Microservices	An architecture principle for cloud based applications, as descibed in the link	https://www.marcobehler.com/guides/java-microservices-a-practical-guide
REST	HTTP based communication between microservices, with structured JSON payload	https://restfulapi.net/
SSE	Server Sent Events: an easy way to transfer small pieces of data between microservices and browsers	https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events
WebSockets	A connection oriented communication, with handshake, data transfer, close, ping and pong	https://websockets.readthedocs.io/en/stable/intro.html https://tools.ietf.org/html/rfc6455
Message Queues	A stateless communication method usually used between one provider and multiple subscribers	https://www.rabbitmq.com/

---

The Environment

• Docker Desktop
  I use the latest version of Docker Desktop (20.10.2) for Windows, installed from the Docker site .
  All containers are connected through a bridge network named "msbasics-network"
• Rabbit MQ
  There is a Docker Official Image for Rabbit MQ. You can get it by calling:
  PS> docker pull rabbitmq
• The Windows command file \init.cmd starts the network and the RabbitMQ container (provided Docker Desktop is running)
  
• for the Java parts, I use OpenJDK 10.0.1 from AdoptOpenJDK
• For the Python software, I use Python 3.9.1
• For C code, I use gcc in two flavors: For Windows, I have MinGW64 from MSYS2, for Linux I use gcc from the Ubuntu release 20.0.4, as a client in Windows WSL2.
• For editing, I use Visual Studio Code.

The Microservices

• API Gateway (port 8080)
  This service handles the interface to the user through port 8080, which is exported from Docker to the outside world. It's internal IP address is 0.0.0.0 since Docker has it's own network and does not know localhost. In the browser, the user calls http://localhost:8080 to get access to API Gateway. The Java Software in API Gateway runs the HTTP Server from com.sum.net.httpserver. The following URIs are served:

  • http://localhost:8080/: The root URI delivers the file index.html on the HTTP GET request.
  • http://localhost:8080/favicon.ico: This URI is called automatically from the broswer to receive a small 16x16 icon.
  • http://localhost:8080/sse: This URI is called automatically by the browser when loading the JavaScript command
    let evtSource = new EventSource("sse/_" + UID);.
    The UID of the Browser window is stored in API Gateway for further handling of SSE Events and when the user closes the browser window.
  • http:localhost:8080/reset: This URI is called from the Browser when loading the AJAX command: $.ajax({ method: "GET", url: "reset" });. The command is called, when the user presses the Reset Button.
  • http://apigateway:8080/summary: The URI is adressed from ServicePython with HTTP POST, thus using the internal docker name of Api Gateway. It delivers the summary value to API Gateway.
  • Independent of these HTTP Services, API Gateway is subscriber to RabbitMQ, with the Queue Name rabbit-msbasics, receiving a counter value every second.

• ServiceJava (port 8081)
  This service generates a running counter and delivers it as provider to RabbitMQ. On the URI http://servicejava:8081/api/v1/ there is a com.sun.net.httpserver that handles a POST command, with the content reset.

• ServiceC (port 8082)
  On the URI http://servicec:8082/api/v1/ there is a http server, derived from the GNU Project libmicrohttpd that handles an incoming PUT request, which contains the reset command and name of a service to be resetted.
  The service also implements two clients based on the native socket.h API:

  • One client is a HTTP client. It triggers the reset command to the named service to receive the reset and transmits the lastcounter, timestamp and casually errormsg data to API Gateway.
  • The other client is a WebSocket client, that transfers the lastcounter and timestamp values to ServicePython.

• ServicePython (port 8083)
  This service runs a WebSocket server on ws://servicepython:8083/, imported from the Python websocket project. On this, it accepts the WebSocket handshake, data transfer and close protocol elements as coming from ServiceC. The received data are lastcounter and timestamp. ServicePython stores up to 10 counter values and delivers the average value as summary via HTTP POST to API Gateway.

Build, Deploy

For the Build+Deploy cycle of every single Microservice, there is a Windows Command file named x.cmd. Before calling any x.cmd you have to start Docker, create the bridge network network-msbasics and start the rabbit-msbasics container. These task are preformed by ìnit.cmd. By calling x.cmd from the respective subdirectory (namely ApiGateway, ServiceJava, ServiceC, ServicePython), the service is built and deployed as a docker container, using the local Dockerfile for the service.
When completed, the Docker Desktop dashboard should look like this:

Now open a browser and go to http://localhost:8080 to see a running counter, a reset button, latest counter and summary.

Copyright, License, Contributing

The code is Copyright (C)2021 by Pius Braun, the License is GPL V3. Feel free to contribute or use parts of it. I would be grateful if you would let me know it.