CLOUD-DIGITAL-LEADER · Question #169
A customer is migrating there on-promises data analytics solution to Google Cloud. The current solution has a lot of data being read form and written to disk. The performance of this approach has occa
The correct answer is C. Use local SSDs with the VMs. Local SSDs are the correct choice to address disk I/O bottlenecks when performance is the priority and the application is fault-tolerant. Local SSDs are physically attached to the VM host and deliver extremely high IOPS (hundreds of thousands) and very low latency compared to net
Question
A customer is migrating there on-promises data analytics solution to Google Cloud. The current solution has a lot of data being read form and written to disk. The performance of this approach has occasionally been a bottleneck for a scale of operations that your cus-tomer has. The application is fault tolerant and can with stand machine going down fre-quently. In moving to Google Cloud they are asking your advice on any way to improve performance?
Options
- AUse Big Query Which has very fast data access and analysis
- BUse Cloud Storage which can be central, scalable storage
- CUse local SSDs with the VMs
- DUse Persistent Disk with the VMs
How the community answered
(55 responses)- A16% (9)
- B7% (4)
- C73% (40)
- D4% (2)
Explanation
Local SSDs are the correct choice to address disk I/O bottlenecks when performance is the priority and the application is fault-tolerant. Local SSDs are physically attached to the VM host and deliver extremely high IOPS (hundreds of thousands) and very low latency compared to network-attached Persistent Disks. Since the application can withstand machine failures (fault tolerant), the non-persistent nature of local SSDs is acceptable - data lost on a VM failure does not break the application. Persistent Disks (Option D) are network-attached and much slower. Cloud Storage (Option B) adds network overhead. BigQuery (Option A) is an analytics service, not a drop-in replacement for a disk-heavy data analytics pipeline.
Topics
Community Discussion
No community discussion yet for this question.