Understanding Serverless: Concepts, Architecture, and Challenges

Serverless is identified by Gartner as the most promising direction for cloud computing in 2019, representing an inevitable trend that reshapes the form of computing resources from the bottom up and introduces new design ideas for software architecture and application deployment.

To help readers master Serverless, a series of articles will be released covering basic concepts, front‑ and back‑end architecture design, application extensions, and best practices.

FaaS vs. IaaS/PaaS – Function‑as‑a‑Service (FaaS) provides a standard runtime and manages request scheduling at the application layer, allowing developers to focus solely on business logic while the cloud provider handles hardware resource maintenance. In contrast, IaaS and PaaS require users to operate virtual machines or container clusters and set up the runtime environment themselves.

How Cloud Functions Solve Problems – Cloud functions (FaaS) form the core of Serverless architecture, handling resource management and request scheduling. Developers write code in a local or web IDE, package it with dependencies, and deploy it as either a function or a service. The platform automatically scales, handles cold‑start mitigation by pre‑warming instances, and predicts concurrency to reserve resources.

Function Shape vs. Service Shape

Service shape supports direct deployment of frameworks such as Node.js Express or Koa without splitting the application into functions, and it can keep a resident instance to reduce cold‑start latency.

Both shapes provide automatic scaling, but function shape creates a new instance per request, leading to consistent performance but potential cold‑start delays when no instances are running.

Serverless 2.0 Component Architecture – The new version enhances product form, system scheduling, and developer tools. It offers local development tools, VS Code extensions, a web IDE, and DevOps integration to improve development, debugging, deployment, and version release workflows. Monitoring and alerting capabilities cover invocation counts, memory usage, concurrency, timeouts, and code errors, addressing infrastructure, resource management, security, and disaster recovery concerns.

Beyond computation, Serverless requires a growing ecosystem, including Serverless DB, performance monitoring, log analysis, orchestration, and high‑performance invocation. Future articles will continue to explore core capabilities to help developers better understand and use Serverless.