Larkspur Systems is building a browser based workspace where teams coauthor very large documents in real time. Each document can reach 25 MB and the system must handle about 60,000 edits per minute…

Question

Options

• ACloud Bigtable
• BCloud Spanner
• CCloud SQL with read replicas
• DFirestore in Datastore mode

• A
  5% (1)
• B
  60% (12)
• C
  25% (5)
• D
  10% (2)

Explanation

Cloud Spanner is the correct choice because it uniquely satisfies all three hard requirements simultaneously: (1) global multi-region replication across all three regions with strong external consistency and serializable transactions - no other managed Google Cloud database offers this combination; (2) it handles millions of reads/writes per second at planetary scale, easily absorbing 60,000 edits per minute; (3) it supports large values and structured data well enough to manage 25 MB documents. Cloud Bigtable (A) offers massive write throughput but only eventual consistency and no ACID transactions, making it unsuitable for a version timeline that requires strong consistency. Cloud SQL with read replicas (C) provides strong consistency but does not scale horizontally across regions without significant manual sharding effort, and read replicas are eventually consistent. Firestore in Datastore mode (D) has a 1 MB per-document limit (documents over that must be split), does not natively guarantee strong consistency across regions in the same way Spanner does, and is better suited for hierarchical document storage rather than real-time collaborative editing at this scale.

No community discussion yet for this question.