Blobs Example

This application demonstrates the use of Blobs that can be used to store and read a blob created by it to and from a storage area respectively.

In this example, the client sends a request to the server which then sends the response back to it. Both the request and response are saved as a blob at the client side to a local file storage at a path given in the configuration.

The client runs on port 9091 and the server on 9090.

Run the example

• Build the application using

  mvn clean package

• Post build use the following commands to:

  • Run the client on localhost:9091

    java -jar target/blobExample-service-1.0-SNAPSHOT.jar sampleClient

  • Run the server on localhost:9090

    java -jar target/blobExample-service-1.0-SNAPSHOT.jar sampleServer

  • Send a request:

    curl http://localhost:9091/displayMessage

    You will know that the app is working fine if you get the correct json response like:

    {
    "clientMessage": "Hello ServerResource. I am ClientResource!",
    "serverResponse": {
    "serverName": "ServerResource",
    "message": "Hi ClientResource! I am ServerResource!"
        }
    }
    

Record Blobs

To enable the recording of the blobs, set enabled to true in config-client.yaml, run the client and server again and hit the above URL.

You will be able to find the blobs created inside blobsRelativePath set in the same config file.

Defaults:

areBlobsEnabled: true

blobsRelativePath: ./blobs

Run Benchmarks

The tool used to benchmark the library is JMH.

Build docker image in local:

docker build -t blobs:blobs-example .

Run docker image

docker run --name blobsExample -v <Source>:/app/bin/blobs_home/JMH-BenchmarkingResults --cpus <Cores> -m <Memory> <IMAGE_ID>

Where,

Source=Full path of the directory you want to save the JMH benchmark run result to.

Cores=Number of CPU cores you want to allocate to the container. Example: 3.0, 5.5, 4.

IMAGE_ID=Docker IMAGE ID which can be found out by running `docker images` in CMD and then looking for `blobs` Repository.

Memory=Container RAM. Example: 8192m.

Look for the json file created in the Source which will contain the results for the benchmark run.

Use JMH-Visualizer(Online Version) to present the benchmark results.

Benchmarking Results

The benchmarking was done for 3 different parameters with the blobs library added to the application to store the blobs in local storage and without storing any blob.

Average Time

Throughput

Sample Time

The benchmarking was performed with following parameters:

• CPU: 4
• RAM: 8GB
• Warm-ups: 10
• Iterations: 15
• Environment: MacOS