BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

AWS Serverless Application Repository To Boost Adoption Of Function As A Service

Following
This article is more than 6 years old.

Amazon announced the general availability of AWS Serverless Application Repository, a marketplace for developers to collaborate on building serverless applications.

Source: AWS

In serverless computing, developers squarely focus on the code and not on the underlying infrastructure. They need not plan the number of servers, amount of storage, and the network topology of deployments. Developers write code snippets in their favorite language and directly upload them to a serverless computing platform for execution. Multiple such code snippets or functions are logically connected to form a complex application. Since the platform deals with one function at a time, and functions are the fundamental deployment units, this model is often called as Functions as a Service (FaaS).

FaaS is fundamentally different from Infrastructure as a Service (IaaS) and Platform as a Service (PaaS). In IaaS, customers are expected to spin up virtual machines to deploy binaries of their application. When dealing with PaaS, developers deploy a codebase of an entire application to the cloud platform. When targeting FaaS, developers don’t think of a whole application. They only write, test, and deploy one function at a time. These functions are assembled together to deliver a unique feature or functionality of an application. So, a function is to FaaS what a VM is to IaaS.

There are three critical attributes to a serverless platform -

  • Per-second billing based on the execution time
  • Transparent resource provisioning through infrastructure abstraction
  • Event-driven invocation

Unlike virtual machines and containers, functions deployed in FaaS are billed only for the execution time. That means when an external application invokes the function, only the duration of execution is considered for billing. Most of the functions running in FaaS are designed to be stateless and short-lived processes. This innovative pricing model makes serverless computing very attractive for certain classes of applications.

AWS took the lead by introducing Lambda as Function as a Service (FaaS) offering. It ushered a new wave of computing by becoming an alternative to virtual machines and containers.

AWS Lambda has come a long way. Multiple AWS services leverage Lambda for processing events. Developers are taking advantage of it for a variety of scenarios including IoT, real-time data processing, chatbots and batch processing.

Though the initial version of Lambda lacked efficient tooling, the recent additions like Serverless Application Model, CodeStar, Cloud9 IDE integration made Lambda accessible to developers.

With Serverless Application Repository, Amazon is taking FaaS to the next level through discoverable, reusable library of applications. It helps developers to clone an existing application template and modifying it to their needs. It has the potential to become a vibrant marketplace of serverless applications. The applications listed in the repository are not just code snippets that run in Lambda. Instead, they are full-blown templates that connect the dots across AWS services. The Serverless Application Model acts the blueprint to bring a variety of assets including S3 buckets, API gateway, SNS and SQL topics, DynamoDB tables and more.

Though it is confined to AWS, Serverless Application Repository can grow to become the Github of FaaS. Eventually, Amazon may enable developers to monetize templates listed in the repository. Ultimately, Serverless Application Repository will drive consumption of Lambda by removing the barriers to embracing serverless computing, which will result in increased adoption of multiple AWS services.

Amazon is continuing to innovate with serverless computing. It is taking advantage of being an early mover in the industry through features like Serverless Application Model and Serverless Application Repository.

Follow me on Twitter or LinkedInCheck out my website