Skip to content

lee212/Serverless-Technologies

Repository files navigation

Latency comparison of serverless computing between IBM Openwhisk and AWS Lambda

AWS Lambda

alt text

  1. As per the use case, I created a trigger from AWS DynamoDB to call AWS Lambda.

  2. The AWS Lambda will then call the lambda function to execute.

  3. In this scenario I had noted down the timestamp of start of the execution of a program that pushes the data to the remote LAMBDA function via AWS dynamodb.

  4. Latency: It is the time an application takes to be on wire. That means from the time the application is triggered from the clients' environemnt(whether UI or code) till the application ends.

  5. I have taken 5 test cases with different input size.The input for each recod remains in same format.Each test pushds a row of data to the dyanamdb table. I captured the latency when the data was transferred from clients' code to update in AWS Dyanamodb database to triggering the AWS Lambda function.Below are the test case and corresponding results

number of records inserted in DB time taken to execute the whole process(in second)
500 1.7 seconds
1000 29.04 seconds
2500 05 min 47.776 seconds
5000 12 mins 38.903 seconds
10000 26 mins 33.789 seconds

IBM Openwhisk

alt text

  1. As per the use case, I created a trigger from IBM cloudant to call IBM openwhisk.

  2. The IBM openwhisk will then call the openwhisk function to execute.

  3. In this scenario I noted down the timestamp of start of the execution of a program that pushes the data to the remote openwhisk function via IBM cloudant

  4. I have taken 5 test cases with different input size.The input for each recod remains in same format.Each test pushds a row of data to the dyanamdb table. I captured the latency when the data was transferred from clients' code to update in IBM cloudant database to triggering the IBM openwhisk function.Below are the test case and corresponding results

number of records inserted in DB time taken to execute the whole process(in seconds)
500 20.7 seconds
1000 1 min 4.675 seconds
2500 04 min 67.569 seconds
5000 11 mins 35.656 seconds
10000 30 mins 23.005 seconds

AWS Lambda Vs IBM Openwhisk

Below is the graph comaprison.As per the chart we can see that IBM Openwhisk-cloudant combination is swift when it comes to small data but AWS lambda-dynamodb combination experience less latency when it comes to upload high amount of data:

alt text

About

Comparison between two Serverless Technologies

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published