Evolution of 'Undifferentiated Heavy Lifting'

It all started with Iaas Providers doing the undifferentiated heavy lifting required for performing mundane activities like spinning up new server instances (Compute), Storage and Networks. Paas provided the deployment and development environments. Thus businesses can focus on their core areas and are not held back by IT/Operations in bringing a new idea quickly to the market. EC2, S3 in AWS or OpenStack Nova, Neutron and Swift can be viewed as 1st Generation of Heavy Lifting.

Next came the concept of Lambda function and we have Serverless, where we can 'pay for value' and AWS Lambda metering done in increments of 100 milliseconds. We have AWS Fargate for running Containers without bothering about servers. Similarly Serverless DBs for Aurora and DynamoDB and even Serverless Orchestration using AWS Step Functions. We can view these as 2nd Generation of Heavy Lifting.

Services like 'AWS Deep Learning Containers', where containers come with pre-installed Docker images are the next generation of Heavy Lifting.They come with deep learning frameworks and simplify deployment of custom Machine Learning (ML) environments. Also with services like 'CodeCommit', 'CodeBuild', 'CodePipeline', 'CodeDeploy', 'CodeStar' and Cloud based IDE like Cloud9, even the development, build and release engineering is heavy lifted. Though these are similar to other Paas providers, but the way they are integrated with different AWS services and charging only for the resources used like EC2 is taking all this to a new level.

Now with all this amount of undifferentiated heavy lifting of Serverless clubbed with Codeless Automation testing, there's nothing holding us back from bringing an idea quickly to market. We are only limited by the number of creative use-cases we can come up with and convert them to business hypotheses. And with these lower and lower barriers of entry, complemented by cheaper broad-band, the world of computing is truly democratic more than ever.