feat(system-design): document replication, sharding, and quorums

[shalshcode08]

Summary

• add a standalone, language-neutral Topic 7 guide to replication, partitioning, sharding, and quorum systems
• cover leader-follower, multi-leader, and leaderless designs; synchronous and asynchronous acknowledgement; failover; conflict handling; read repair; anti-entropy; and hinted handoff
• explain strict and sloppy quorums, replica placement, range and hash partitioning, consistent hashing, virtual nodes, partition maps, fan-out, hotspots, tenant placement, and online resharding
• compare sequence allocation, UUID variants, and Snowflake-style identifiers, including ordering leakage and clock risks
• include a complete sharded order design with replica and network failure timelines, capacity estimates, operations, security, testing, exercises, interview answers, and a revision cheat sheet
• register the new page in the System Design navigation without changing styles, shared design, tests, or TOPICS.md

Research

Primary official sources include PostgreSQL current replication and table-partitioning documentation, Apache Cassandra architecture and repair documentation, MongoDB sharding and balancer documentation, the Amazon Dynamo paper, RFC 9562, and Google Cloud Spanner schema-design guidance.

Validation

• npx prettier --write manifest.json src/content/systemdesign/replication-partitioning-sharding-quorum-systems.html
• npx html-validate src/content/systemdesign/replication-partitioning-sharding-quorum-systems.html
• npm test - 22 of 22 passing
• npm run build - 37 pages built successfully
• git diff --check
• exact footer verification
• no em dash scan
• scope verification: only manifest.json and the Topic 7 content page

Constraints preserved

• one topic only
• no CSS or shared design changes
• no test changes
• no TOPICS.md status change before merge
• exact July 2026 documentation footer included

Review focus

Please focus on quorum caveats, online migration safety, failure-domain assumptions, identifier clock risks, and whether the order-system decisions remain explicit about workload trade-offs.