How AWS powers Delivery Hero's Logistics Optimization Process
summary
Delivery Hero runs a distributed logistics stack on AWS that performs a 1-minute global optimization cycle, dispatching millions of orders daily to 500k+ active couriers. Architecture uses EC2/EKS (about 1,000 cores and 4 TB RAM claimed), 70% spot instances, SNS/SQS messaging with peaks around 450k messages/min, and ElastiCache (Redis) serving 2.6M GETs/min for recoverable, low‑latency state. BRP routing algorithms run in containers, DynamoDB/RDS store durable configs and S3 stores per‑cycle snapshots; the system scales and tolerates failures in production.
AI-written summary · about
chapters (8)
- 0:00Introduction and context
- 1:53Logistics optimization scope
- 6:17Architectural principles
- 11:16AWS infrastructure overview
- 16:25Compute and messaging layers
- 21:53Data and state management
- 25:46High-level architecture view
- 30:21Learnings and best practices
related talks
31mDevOps in the Cloud: Case Studies of Amazon.com teams and their resilient architectures
52mHow rapid growth accelerated Zipwhip's move to AWS
51mArchitecting for Hyperscale with AWS: Navigating Real-world Serverless Challenges
29mHow local businesses are rapidly achieving global scale on AWS
29mAccelerate Innovation with Hybrid Capabilities and Infrastructures
22mFireside Chat with AWS
01h 10mHosting Drupal on AWS
23mOn-demand Container Loading in AWS Lambda