AdvancedSystem-Design
45 min
Design a Cloud Data Warehouse
Data WarehouseStorageETLAnalytics
Advertisement
Interview Question
Design a data warehouse platform supporting petabyte-scale storage, ELT/ETL pipelines, query federation, and cost controls.
Key Points to Cover
- Columnar storage and MPP architecture
- ETL/ELT pipelines, batch vs streaming loads
- Query federation and optimizer design
- Concurrency scaling, workload management
- Governance: lineage, RBAC, cost monitoring
Evaluation Rubric
Efficient columnar storage & MPP25% weight
Robust ETL/ELT design25% weight
Strong federation & optimizer plan25% weight
Governance & cost management25% weight
Hints
- 💡Leverage separation of storage and compute.
Common Pitfalls to Avoid
- ⚠️Underestimating the complexity and cost of data governance and security at petabyte scale.
- ⚠️Failing to adequately plan for data volume growth and scalability limitations of chosen technologies.
- ⚠️Over-reliance on a single ETL/ELT tool without considering diverse source system requirements.
- ⚠️Neglecting query optimization and performance tuning, leading to high compute costs and slow analytics.
- ⚠️Lack of robust monitoring and alerting, resulting in undetected performance degradations or cost overruns.
Potential Follow-up Questions
- ❓How do you handle schema evolution?
- ❓How to isolate workloads between teams?
Advertisement