AWS Lambda: Four Compute Primitives

One service name, four execution models with different units, durations, and isolation

AWS Lambda: Four Compute Primitives One service name, four execution models with different units, durations, and isolation One service, four execution models AWS Lambda · four compute primitives · Architecture component AWS Lambda four compute primitives Lambda Functions · per-invocation, 15 min max · One service, four execution models Lambda Functions per-invocation, 15 min max Managed Instances · customer EC2, always-on, parallel · One service, four execution models Managed Instances customer EC2, always-on, parallel Durable Functions · checkpoint / replay, up to 1 year · One service, four execution models Durable Functions checkpoint / replay, up to 1 year MicroVMs · Firecracker direct, 8 h, snapshot · One service, four execution models MicroVMs Firecracker direct, 8 h, snapshot Event Sources · API GW, SQS, Kinesis, schedules · Architecture component Event Sources API GW, SQS, Kinesis, schedules Direct Clients · HTTPS / WebSocket / gRPC · Architecture component Direct Clients HTTPS / WebSocket / gRPC default code on your EC2 Functions + checkpoints VM-level sandbox invoke per-VM HTTPS endpoint Legend Backend Cloud External

Architecture diagram • Built with Archify • Create yours ↗ • Hover to trace • R route • Click to focus • +/− zoom • M radar • T theme • E export