Your application is controlled by a managed instance group.
You want to share a large read-only data set between all the instances in the managed instance group.
You want to ensure that each instance can start quickly and can access the data set via its filesystem with very low latency.
You also want to minimize the total cost of the solution.
What should you do?
Click on the arrows to vote for the correct answer
A. B. C. D. E.C.
To share a large read-only data set between instances in a managed instance group, there are a few options available. The solution should ensure that each instance can start quickly and access the data set via its filesystem with very low latency while minimizing the total cost of the solution. Let's examine the possible answers:
A. Move the data to a Cloud Storage bucket, and mount the bucket on the filesystem using Cloud Storage FUS.
This solution involves moving the data set to a Cloud Storage bucket and using Cloud Storage FUSE to mount the bucket on the filesystem of each instance in the managed instance group. FUSE stands for Filesystem in Userspace, which means that the filesystem is implemented in user-level software rather than in the operating system kernel. This approach can be beneficial because it allows you to avoid copying the data set to each instance's local disk, which can save storage space and reduce boot times. However, there may be some performance overhead with using FUSE, which could impact latency.
B. (No answer given)
C. Move the data to a Cloud Storage bucket, and copy the data to the boot disk of the instance via a startup script.
This solution involves moving the data set to a Cloud Storage bucket and copying the data to the boot disk of each instance in the managed instance group via a startup script. This approach can be beneficial because it allows you to ensure that each instance has a local copy of the data set on its disk, which can reduce latency since the data is stored locally. However, there may be some additional time required during instance boot-up to copy the data from the bucket, which could impact startup time.
D. Move the data to a Compute Engine persistent disk, and attach the disk in read-only mode to multiple Compute Engine virtual machine instances.
This solution involves moving the data set to a Compute Engine persistent disk and attaching the disk in read-only mode to each instance in the managed instance group. This approach can be beneficial because it allows you to ensure that each instance has a local copy of the data set on its disk, which can reduce latency since the data is stored locally. Additionally, because the disk is read-only, it can be shared among multiple instances, reducing the storage cost. However, there may be some additional time required during instance boot-up to attach the disk, which could impact startup time.
E. Move the data to a Compute Engine persistent disk, take a snapshot, create multiple disks from the snapshot, and attach each disk to its own instance.
This solution involves moving the data set to a Compute Engine persistent disk, taking a snapshot of the disk, and creating multiple disks from the snapshot. Each disk is then attached to its own instance in the managed instance group. This approach can be beneficial because it allows you to ensure that each instance has a local copy of the data set on its disk, which can reduce latency since the data is stored locally. Additionally, because each disk is independent, it can be scaled individually, reducing the storage cost. However, there may be some additional time required during instance boot-up to attach the disk, which could impact startup time.
Overall, the best solution will depend on the specific requirements of the application. If low latency is a priority, options C, D, and E may be preferable, as they all allow for local storage of the data set. However, if minimizing storage cost is a priority, options A and D may be preferable, as they allow for shared storage of the data set.