From Satellites to Slobs
summary
Overview of the pipeline from raw satellite imagery through gal-based tiling into user maps and the infra choices that determine speed cost and reliability: Docker containers package gal and tiling tools, Kubernetes enables horizontal scaling but runs roughly 5× the cost of serverless for this use case, and serverless hosts full tiling containers on demand cutting monthly bills dramatically while introducing cold‑start latency. Processing a single high‑resolution zoom level can take days to months, Cloudflare caching and monitoring prove essential and the serverless tiling approach works with caveats.
AI-written summary · about
chapters (8)
- 0:00Introduction to mapping software
- 2:34Abstraction in software design
- 5:43Tiling maps for web delivery
- 8:09Backend infrastructure for scaling
- 11:22Cost implications of scaling
- 15:59Serverless vs container costs
- 18:45Frontend mapping technologies
- 19:26Q&A on infrastructure costs
related talks
19mEstablishing a Serverless Dynamic Tile Generation Workflow
50mBehind the scenes of Mapbox Satellite imagery
37mMosaicking the Earth every day
25mA Journey Through a Geospatial Data Pipeline: From Raw Coordinates to Actionable Insights
23mFrom field to screen: Designing for real-time conservation impact
26mWho Said Wrangling Geospatial Data at Scale was Easy?
31mOpenDataCube - Petabyte Scale Satellite Imagery Processing in Python
05mContainerization Beyond CPUs – A Kubernetes-based Serverless Pl...