Real-time sports scores with low-latency updates, fanout, caching, and resilience.
Build a reliable ingest path with ordering, dedupe, and a durable event log you can replay.
What was missing: Durability, ordering, and replay for partner events.
Why that's risky: Without an append-only log, bugs/outages permanently lose or corrupt score updates.
What gets added: Edge/WAF, partner ingestion API, event stream (partitioned by match), event log retention, snapshot builder + snapshot DB.
Trade-offs: Eventual consistency between ingest and snapshot read model; operational overhead of streaming.
Serve millions of reads safely using CDN + cache; keep DB off the critical fan traffic path.
What was missing: Read scalability.
Why that's risky: Finals traffic melts the DB and increases tail latency.
What gets added: CDN in front of read API + hot cache; snapshot DB as the source of truth for reads.
Trade-offs: Cache invalidation.
Fan out ordered deltas to connected clients using realtime gateways and a pub/sub layer.
What was missing: Realtime delivery.
Why that's risky: Polling at scale is expensive, laggy, and causes thundering herds.
What gets added: Realtime LB + gateways, delta publisher, and a pub/sub layer for efficient fanout.
Trade-offs: Connection management + backpressure; ordering guarantees per match require careful partitioning.
Make the system resilient: retries/DLQ, observability, and multi-region failover for big matches.
What was missing: Operational safety and regional resilience.
Why that's risky: One region outage or a bad partner event can cause global downtime or incorrect scores.
What gets added: Multi-region deploy, regional caches, DB replicas, retry/DLQ path, and observability with SLO alerts.
Trade-offs: Cost + complexity of multi-region replication and operational tooling.
Write compensating events with higher sequence/version; consumers rebuild snapshots; clients receive correction deltas.
Validate at ingestion, quarantine failures into a DLQ, and use delayed retries with alerting. Keep the stream moving by skipping/isolating bad partitions and replay once fixed.
You're in the middle of an interview session. Leaving now will end your current attempt.
Explore concept overviews, real-system examples, key tradeoffs, and interview talking points for each roadmap section.
You've conquered this phase. These are the skills you now own: