In the ever-evolving landscape of computing, serverless has undeniably established itself as a central pillar. The driving force behind this transition is the growing availability of serverless offerings from major cloud providers such as Amazon Web Services (AWS), Google Cloud, and Azure, along with emerging platforms like Vercel and Cloudflare.
This report provides a comprehensive analysis of how over 20,000 organizations are utilizing serverless technologies in their operations, exploring significant trends and insights drawn from real-world applications of this transformative technology.
Shift Toward Serverless Adoption
Significant growth has been observed in serverless adoption among organizations operating on Azure and Google Cloud, with AWS also showing positive development. For instance, 70% of the AWS customers and 60% of Google Cloud customers now use serverless solutions. Azure isn’t far behind, with 49% of its customers embracing serverless offerings.
This upswing can be attributed to the expanding suite of serverless tools, ranging from FaaS solutions to serverless edge computing, offered by these cloud providers to meet their customers’ unique needs.
The Rise of Container-Based Serverless Computing
Google Cloud, since its launch of Cloud Run in 2019, has led in fully managed container-based serverless adoption. However, this year AWS saw a rise to 26% of serverless organizations running containerized Lambda functions and AWS App Runner. Azure also experienced considerable year-over-year growth, propelled by the launch of Azure Container Apps.
Container-based serverless compute platforms are gaining traction as they facilitate serverless adoption and migration by enabling organizations to deploy existing container images as microservices. Apart from that, these platforms offer wider language support and larger application sizes.
Serverless Platforms: Beyond The Major Providers
While major providers dominate the serverless space, frontend development and Content Delivery Network (CDN) platforms like Vercel, Netlify, Cloudflare, and Fastly also equip developers with specialized serverless compute capabilities. Interestingly, 7% of organizations monitoring serverless workloads in a significant cloud are also running workloads on one or more of these emerging platforms.
Choice of Languages for AWS Lambda
Node.js and Python are the languages of choice for most AWS Lambda developers, with over half of invocations being written in these languages. The rising popularity of custom runtimes indicates a growing interest in serverless containers, which allow developers to work with languages not natively supported by Lambda.
The Challenge of Cold Starts
Cold starts, where a new execution environment is created to serve a request, remain a significant concern. This is especially true for Java-based Lambda functions, which showcase the longest cold start times due to the JVM and Java libraries’ loading time.
The Adoption of AWS Lambda on ARM
The usage of AWS Lambda on ARM has doubled in the past year, primarily due to its combined benefits of faster execution times and lower costs.
Deployment Tools for AWS Lambda
Infrastructure as Code (IaC) tools like the Serverless Framework and Terraform greatly simplify the deployment and configuration of Lambda functions and other resources. As organizations mature and scale, the preference for IaC tools shifts. Larger organizations positively inclined towards Terraform for multi-cloud support and flexibility.
Connection of AWS Lambdas to a Virtual Private Cloud (VPC)
The complexity of integrating serverless functions across the existing infrastructure has led many organizations to connect their Lambda functions directly to the VPCs. According to recent statistics, 65% of Datadog customers have at least five Lambda functions connected to a dedicated VPC in their AWS account.
Serverless technologies today are making developer’s lives easier by being more secure, cost-effective, flexible, and efficient. The prominence of serverless in modern application building is only expected to surge further in the coming years.