Which Storage Option Gives Highest Throughput for Hundreds of VMs?

PCA Last updated

Configuring individual storage systems. Designing network, storage, and compute resources.

Short answer

B Create a Filestore Enterprise instance in the same region and mount its file share on every Compute Engine VM for maximum shared throughput.

Question

A large enterprise is building a file processing pipeline on Google Cloud. The pipeline ingests the raw files, which are then accessed by a fleet of hundreds of Compute Engine virtual machines (VMs) in a single region for parallel processing. These VMs need concurrent, low-latency read and write access to a shared file system. You need to implement a storage solution that delivers the highest possible throughput to minimize the overall processing time. What should you do?

  1. Use a multi-region Cloud Storage bucket to store the files.
  2. Create a Filestore Enterprise instance in the same region as the VMs. Mount the Filestore file share on each Compute Engine VM. Correct Answer
  3. Create a Filestore Zonal instance in each zone where the VMs are running. Mount the Filestore file share on each Compute Engine VM.
  4. Create a Regional Persistent Disk and attach it to all VMs in multi-writer mode.

Key Takeaways

The test is whether you can distinguish one genuinely shared, regional NFS file system from object storage, per-zone isolated shares, or a block disk used as shared storage — and the trap is picking a Zonal Filestore per zone.

This PCA question asks which Google Cloud storage design gives hundreds of Compute Engine VMs concurrent, low-latency, high-throughput shared file access in one region. The verdict is a single Filestore Enterprise instance created in the same region as the VMs, with its NFS file share mounted on every VM (B).

Common mistake: Option C (a Filestore Zonal instance in each zone) is the most-chosen wrong answer: Zonal instances are independent and do not replicate, so the fleet would end up with several unrelated file shares and would have to partition or sync the raw files itself, breaking the single shared file system requirement.

Expert Analysis

Why the Answer Is Correct

Filestore Enterprise creates a regional, multi-zone NFS file system, so a single file share can be mounted by every Compute Engine VM in the region regardless of zone, giving the pipeline one true shared namespace. Enterprise tier is designed for high-performance enterprise workloads and delivers the highest per-instance throughput and IOPS of the Filestore tiers, which is exactly what minimizes total processing time for hundreds of parallel readers and writers. Co-locating the instance in the same region as the VMs keeps network latency low, and NFS natively allows many clients to read and write concurrently without any application-side synchronization.

Official Reference

Exam Strategy

Always check whether the scenario requires one shared namespace before comparing performance numbers — a tier that is faster in isolation but creates isolated shares cannot satisfy a shared file system requirement. Match the tier (Standard, Zonal, Regional, Enterprise) to both the latency zone scope and the throughput ceiling the question emphasizes.

Frequently Asked Questions

Why is a Filestore Zonal instance in each zone wrong if it has low latency?

Zonal instances are independent and do not replicate data, so the VMs would see several separate file shares instead of one shared file system, forcing manual partitioning or syncing of the raw files.

Why not use a multi-region Cloud Storage bucket for the pipeline?

Cloud Storage is object storage with no mounted POSIX file share and no low-latency concurrent read/write semantics for hundreds of VMs, so it cannot serve as the shared file system the pipeline requires.

More PCA FAQ →

Discussion 2 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded.

From public community threads 2

  1. SCSK 👍 1 Answered C
    Filestore Zonal está diseñado para cargas de alto rendimiento y baja latencia. Crear una instancia en cada zona permite mantener el almacenamiento cerca de las VMs consumidoras y distribuir el procesamiento entre instancias zonales. El acceso se realiza mediante NFS, permitiendo que múltiples VMs lean y escriban simultáneamente en un sistema de archivos compartido.
  2. Jucara 👍 1
    C is incorrect: Creating a Filestore Zonal instance in each zone creates isolated, independent file shares that do not replicate data between them. The ingestion pipeline would have to manually partition or synchronize raw files across multiple instances, violating the requirement for a single shared file system.

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Submitted for moderation before publishing. Keep it helpful and respectful.

Optimizing Data Transfer for ETL with Cellular Connections

S-Grade

Streamlining Cloud IAM Audit Analysis with BigQuery

S-Grade

Auditing BigQuery User Query Counts

S-Grade

Practice All PCA Questions

Access 81 questions with complete answers and detailed explanations.

View Full PCA Practice Test →

← Back to PCA Study Guide