A mini project made to implement concepts of client server architecture.The distribution of load takes place based on the CPU usage of each worker nodes.The client sends a number to the server and it receives the prime number which are smaller that the provided number.Ex:Input: 100Output: 97
Input: 6Output: 5
Client will request use for server's IP and connect to the server on port 7777.After that the server can proceed to provide inputs. Typedone when user is willing to quit the application.
Server listens at port 2222 for the worker nodes to connect. Large number of worker nodes are supported at anytime, which makes this project highly scalable.Server also listens on port 7777 for client requests. docker-compose will bind the localhost:7777 to server's 7777 port. Server needs no input from the user.
The worker node requests a connection from server at port 2222 and it's always listening on port 5555.Following which server establishes a TCP connection to the workers on port 5555. Worker nodes need no input from the user.
It requiresjava,docker anddocker-compose to work.Please follow the installation process from the official documentation of all the dependencies.
UseDockerfile in server and farm to build an image.
cd loadBalancerdocker build -t"loadbalancer/server" Server
Build Worker image:
docker build -t"loadbalancer/worker" Worker
Once images are created we will use docker-compose to turn up the server and the worker nodes
Now we can use client to connect to the server
cd loadBalancer/Client/javac ClientMain.javajava ClientMain
Provide localhost as the server IP if containers are also running on the same machine.If you are remotely connecting to the server then provide the IP of host which is running the server-worker containers.
Inputdone when you want to quit.
- Fix the server crash when clients exit.
- Support to very large number inputs.