Cloud Drive 2 - Enterprise Collaboration & Compliance

Interview-Ready Approach

1) Clarify Scope and SLOs

• •Problem statement: Scale to 100 M users with real-time co-editing, eDiscovery, data loss prevention, and multi-region replication.
• •Design for a peak load target around 6,944 RPS (including burst headroom).
• •Total users: 100,000,000
• •Total storage: ~500 PB
• •Concurrent editors (per doc): Up to 100
• •Collaboration sync latency: < 200 ms
• •Search latency: < 1 second

2) Capacity Planning Method

• •Convert traffic and growth constraints into request rate, storage growth, and concurrency budgets.
• •Keep at least 2-3x safety margin per tier (ingress, compute, storage, async workers).
• •Reserve explicit latency budgets per hop so p95 can be defended in review.

3) Architecture Decisions

• •Storage: Use object storage for large blobs and keep metadata/authorization separate in the API tier.
• •Databases: Define a clear system-of-record and design read/write paths separately before adding optimizations.
• •Search: Use primary store for writes and async index updates for search relevance + scale.
• •Auth: Centralize identity verification and keep authorization checks close to domain resources.
• •Monitoring: Instrument golden signals (latency, traffic, errors, saturation) per tier and per tenant/domain.
• •Replication: Separate primary write path from replicated read path and define lag tolerance per feature.

4) Reliability and Failure Strategy

• •Enforce lifecycle policies, retention tiers, and checksum validation.
• •Use strong write constraints (transactions or conditional writes) and explicit backup/restore strategy.
• •Track indexing lag and support reindex from source of truth.
• •Use short-lived tokens and secure key rotation workflows.
• •Alert on user-impact SLOs, not only infrastructure metrics.

5) Validation Plan

• •Run one peak-load test, one dependency-degradation test, and one failover test.
• •Verify idempotency for all retried writes and async consumers.
• •Track user-facing SLOs first: p95 latency, error rate, and successful throughput.

6) Trade-offs to Call Out in Interviews

• •Storage: Object storage is cheap and durable, but random low-latency reads are weaker than databases/caches.
• •Databases: SQL gives stronger transactional guarantees; NoSQL often gives better write scaling and flexibility.
• •Search: Search index gives rich querying but introduces eventual consistency and index ops overhead.
• •Auth: Central auth simplifies policy, but makes auth service availability/security critical.
• •Monitoring: Deep observability speeds incident response but raises ingestion and tooling costs.

Practical Notes

• •CRDTs (Conflict-free Replicated Data Types) are simpler to reason about than OT for collaborative editing at scale.
• •Index file contents with Elasticsearch - use Apache Tika to extract text from PDFs, DOCX, etc.
• •DLP scanning can be an async pipeline - upload → object store → DLP scanner → flag/approve → make available.

Why This Solution Works

Request path: The solution keeps ingress, service logic, and stateful dependencies separated so each layer can scale independently.

Reference flow: Web Clients -> Load Balancer -> API Gateway -> Core Service -> Auth Service -> Primary SQL DB -> Read Model DB -> Object Storage

Design strengths

• •Monitoring and logs are wired in from day one for rapid incident triage.
• •Security controls are enforced at ingress to protect downstream capacity.

Interview defense

• •This design makes bottlenecks explicit (ingress, core compute, persistence, async workers).
• •It supports progressive scaling without re-architecting the core request path.
• •It keeps correctness-sensitive state changes in durable systems while offloading background work asynchronously.