HLDArena
HLD-318·Geo & Telemetry
Hard · 1,500 pts

Live Road-Segment Traffic Telemetry & Partitioned ETA Engine

Design Uber / Google Maps’ real-time ETA engine ingesting continuous driver speed probes onto partitioned road-graph tiles and serving low-latency route ETAs.

Max p99
≤ 45 ms
Availability
≥ 99.5%
Budget
≤ $8,500

Functional Requirements

  • FR-1POST /traffic/probes — Aggregates live GPS speed telemetry onto road segment tile_id weights.
  • FR-2GET /eta/compute?tile={tile_id} — Evaluates contraction-hierarchy / road-tile edge weights under 45ms p99.
  • FR-3DELETE /traffic/probes/{tile_id} — Supports tenant retention TTL expiry, rule updates, or stream cleanup while keeping recent rollup caches fresh.
  • FR-4Must sustain rush-hour traffic storms and survive primary road-state shard crashes.

Workload Profile

Midday Traffic50,000 ETA Read RPS / 15,000 Probe Write RPS
Evening Rush Hour Peak135,000 Read RPS / 34,000 Probe Write RPS
Active Working Set32 GB metropolitan road-graph edge speed tables
Traffic DistributionDense spatial skew on downtown & airport corridor tile_id shards

Consistency Model

Bounded eventual road-weight updates (< 5s traffic lag) cached in regional tile memory.

Design Write-up
0 boxes · 0 edges
0 words
Hidden Cases #4–#7 run automatically on Submit
Case 1: Smoke & API ContractInput Workload
Read RPS
2,000 rps
Write RPS
400 rps
Key Distribution
Uniform (Low Skew)
Chaos Fault
None (Healthy Cluster)
Expected Pass CriteriaASSERTIONS
• Tail Latency: p99 ≤ 90 ms
• Availability: ≥ 99.5%
Valid end-to-end Client → Compute → Storage topology