Examkin

Question

Difficulty: HardPlanning Database Options and Persistent Disk Types

A software company is architecting a custom high-throughput transaction processing engine deployed on Compute Engine virtual machines. The engine requires ultra-low latency block storage for temporary operational scratch data during execution, but requires durable storage for its primary relational data store with continuous low latency, high random IOPS, availability across VM restarts, and support for automated disk snapshots. Which storage configuration meets these requirements according to Google Cloud best practices?

  1. Attach a Local SSD array for temporary operational scratch data, and use SSD Persistent Disk (pd-ssd) for the primary relational data store.Answer
  2. B
    Attach Local SSD for both the temporary scratch data and primary data store, using Spot VMs to lower operational costs.
  3. C
    Deploy Cloud Bigtable as the primary data store for relational transaction processing to eliminate block storage overhead.
  4. D
    Use Standard Persistent Disk (pd-standard) for the primary data store and move disk snapshot files to Cloud Storage Archive class daily.

Answer

Attaching a Local SSD for temporary scratch space and using SSD Persistent Disk (pd-ssd) for the primary relational data store provides the required low latency, high IOPS, persistence across reboots, and snapshot functionality.
Combining Local SSD for non-persistent scratch workspace with SSD Persistent Disk (pd-ssd) for the persistent database volume balances extreme performance with enterprise durability and backup capabilities. SSD Persistent Disk ensures high IOPS, durability across VM restarts, and point-in-time snapshot features.

Step-by-Step Solution

1
Analyze storage performance and durability requirements for scratch space.
Temporary scratch data requires maximum IOPS and ultra-low latency, which Local SSD provides directly attached to the host server.
Local SSD delivers high throughput and microsecond latency, but loses data on instance termination/stop.
2
Analyze storage durability, snapshot, and IOPS requirements for primary relational data.
SSD Persistent Disk (pd-ssd) offers high performance, persistence across instance restarts, and support for automated snapshot backups.
Persistent disks are network-attached and support snapshot backups and persistent lifecycle independent of VM state.
3
Evaluate and eliminate unsuitable storage types.
Local SSD cannot back primary persistent database files because it lacks snapshot capability and persistence. Cloud Bigtable is NoSQL rather than relational. Standard persistent disks lack performance for databases.
Matching storage types strictly to data lifecycle and workload characteristics ensures architectural reliability and performance.

Key Concept

Selecting GCP Persistent Disk types and managed storage options based on IOPS, latency, durability, and workload characteristics.
Rate this question

Topics

Planning Database Options and Persistent Disk Types

More questions from this topic