10 of 10 cards free · no sign-in needed
General Purpose (balanced), Compute Optimised (CPU-intensive), Memory Optimised (large in-memory workloads), Storage Optimised (high I/O), Accelerated Computing (GPU/FPGA)
On-Demand: pay per second, no commitment. Reserved: 1 or 3-year commitment, up to 72% discount. Spot: unused capacity, up to 90% discount, can be reclaimed with 2-min notice.
Serverless compute — run code in response to events (API calls, S3 uploads, DynamoDB changes) without managing servers. Pay only for execution time (per millisecond). Auto-scales instantly.
Automatically adds (scale out) or removes (scale in) EC2 instances based on demand metrics like CPU or request count. Ensures right capacity at right time without manual intervention.
Distributes incoming traffic across multiple EC2 instances. Routes traffic only to healthy instances. Enables horizontal scaling and high availability. Works with Auto Scaling.
ECS: AWS-managed container orchestration, simpler to operate. EKS: Managed Kubernetes (industry-standard open-source). Both support EC2 and Fargate launch types.
Serverless compute engine for containers (ECS and EKS). No EC2 instances to manage — define CPU/memory requirements and Fargate handles the rest. Pay only for resources used.
Pub/Sub messaging: publishers send to a topic, subscribers (email, SMS, Lambda, SQS, HTTP) receive notifications. Used for fan-out patterns and event-driven architectures.
Managed message queue for decoupling components. Messages persist until consumed. Supports standard (best-effort ordering) and FIFO (exactly-once, ordered) queues.
Sign in (free) to flip every card, mark what you know, and pick up exactly where you left off.
After sign-in: Space flip · ←→ navigate · K knew it · S skip