Tim's Wiki

Serverless

Properties2
descriptionApplication architecture approach that uses managed services and event-driven functions so teams can ship without operating traditional always-on servers.
tagstech, cloud, mature

2 min read

Serverless

Definition

Serverless is an application architecture style that reduces or removes the need to manage traditional always-on servers directly. It typically combines managed cloud services with event-driven compute, especially Functions as a Service (FaaS), so teams focus more on business logic and less on infrastructure operations.

Core Ideas

Main Building Blocks

  • FaaS — run code in short-lived execution environments, such as AWS Lambda
  • BaaS — use managed backend services like authentication, storage, queues, and databases
  • Event-driven execution — functions are triggered by HTTP requests, queue messages, file uploads, schedules, or streams
  • Managed orchestration — workflows can be coordinated with services like AWS Step Functions

Typical Benefits

  • Lower operational overhead — no server patching, fleet management, or manual scaling
  • Elastic scaling — capacity grows and shrinks with incoming events
  • Faster iteration — small deployable functions support rapid changes
  • Cost alignment with usage — many platforms charge per invocation or execution time rather than fixed capacity

Common Trade-Offs

  • Vendor dependence — architecture often becomes tied to a cloud provider’s primitives
  • Cold starts — sporadic workloads can pay a latency penalty when functions spin up
  • Observability complexity — logs, traces, and debugging become harder across many event boundaries
  • Execution constraints — short runtimes, stateless execution, and platform limits shape system design

Common AWS Patterns

  • API Gateway + Lambda for HTTP APIs
  • S3 event notifications triggering processing workflows
  • SQS/SNS/EventBridge for decoupled event handling
  • Step Functions for multi-step workflows, retries, and branching
  • Kinesis for ingestion and stream-oriented data processing

When It Fits Well

  • bursty or unpredictable workloads
  • event-driven automation
  • lightweight APIs and internal tools
  • asynchronous data processing pipelines
  • teams that want to avoid managing cluster or server infrastructure

When It Fits Poorly

  • long-running or stateful workloads
  • latency-sensitive systems where cold starts are unacceptable
  • platforms with heavy local resource requirements
  • workloads where provider lock-in is strategically unacceptable

Relationships

  • Cloud & AWS Infrastructure — cloud services provide the building blocks for serverless architectures
  • RESTful API — many serverless systems expose HTTP APIs through API gateways
  • Microservices — serverless can be used as an execution model for small independently deployed services
  • Software Architecture & Distributed Systems — serverless is one of the major distributed application styles in the source notes
  • Cloudflare OS — a platform built entirely on serverless primitives (Workers, Durable Objects, R2), and the lock-in that creates

References

  • _Serverless Study Index
  • AWS Serverless Course
  • 7 Serverless

Graph View

Backlinks