Glossary · devops
Docker
Definition
Docker is a platform that packages an app and its dependencies into an isolated 'container'. It solves 'works on my machine' — the same container runs identically on laptop, server, cloud. It's been DevOps's core infrastructure since 2013.
Detailed explanation
Container vs VM: VM bundles a full OS (GBs, minutes to boot); container shares the OS kernel + only app + lib (MBs, seconds to boot). 100 containers per server vs 5-10 VMs.
Components: Dockerfile (build instructions), Image (runnable package), Container (running instance), Registry (Docker Hub, AWS ECR, GitHub Container Registry — image storage).
Compose + Kubernetes: Docker Compose for small dev/prod stacks (docker-compose.yml). Kubernetes for 10+ container production orchestration. ECS (AWS) + Cloud Run (Google) are managed alternatives.
2026 trends: distroless images (reduced attack surface), multi-stage builds (smaller images), BuildKit (cache + parallel), Buildpacks (Dockerfile-free builds).
Use cases
→Standardize developer environments
→CI/CD pipeline (test + build + deploy)
→Microservice distribution
→ML/AI inference deployment
→Multi-region production
Pros
- +Solves 'works on my machine'
- +Fast boot (seconds)
- +Resource-efficient (100 containers/server)
- +Standard format (Docker Hub ecosystem)
Cons
- −Networking + volume complexity
- −Docker Desktop license (enterprise)
- −Image security (CVE scanning required)
- −Kubernetes adds learning curve in production
Related terms
Related services
Planning a project around Docker?
In a 30-minute discovery call we share a written architecture + cost + team recommendation tailored to your project.
Start a discovery call