#16Distributed Web Crawler
16 / 20
Google logoGoogle System
HardDistributed Systems
40–60 min

Distributed Web Crawler

Architect a scalable web crawler that downloads, parses, deduplicates, and indexes billions of web pages while strictly respecting robots.txt and politeness policies.

Scale & Traffic10 Billion pages/month · 5,000 pages/sec download rate
Monthly Cloud Budget$25,000/month
Latency Target (P99)< 50ms Realtime SLA
Availability SLA99.99% Multi-AZ Active

System Requirements

  • Frontier URL manager ensuring host-level politeness delays
  • Content deduplication using 64-bit SimHash / MinHash fingerprints
  • DNS caching layer to eliminate domain resolution bottlenecks
  • Raw HTML blob storage + extracted link graph pipeline

Graded Constraints

01Budget: $25,000/month
02Distributed priority queue for URL frontier management
03Worker cluster for fetching and parsing HTML
04DNS cache server cluster
05Object storage for raw HTML snapshots
06NoSQL database for crawled metadata and document checksums

Ready to design?

Open this challenge in the interactive system design playground and satisfy all architecture constraints.

Explore More

Related Architecture Challenges

All Problems
Distributed Web Crawler — System Design Problem | Mimir Nest