GCP Persistent Disk & Filestore: Block and File Storage
TL;DR
GCP Persistent Disk offers the industry’s most cost-effective block storage at $0.048/GB for SSD. Filestore provides managed NFS at $0.20/GB — cheaper than AWS EFS. Together they cover the storage spectrum: PD for VM disks, Filestore for shared file systems. Key differentiator: PD’s live snapshotting and global image sharing. The catch: smaller feature set than AWS EBS/EFS and GCP-only.
What Is It?
Persistent Disk (Block Storage)
Durable, high-performance block storage for Compute Engine VMs.
| Type | Price/GB | Max IOPS | Use Case |
|---|---|---|---|
| Standard | $0.040 | 7,500 | Bulk storage |
| SSD (pd-ssd) | $0.048 | 100,000 | General purpose |
| Balanced | $0.060 | 80,000 | Price/performance |
| Extreme | $0.125 | 350,000 | Databases |
Filestore (Managed NFS)
Fully managed NFS for shared file systems.
| Tier | Price/GB | Throughput |
|---|---|---|
| Basic HDD | $0.16 | 100 MB/s |
| Basic SSD | $0.20 | 1 GB/s |
| High Scale | $0.40 | 10+ GB/s |
| Enterprise | $0.40 | 25 GB/s |
Architecture Patterns
Pattern 1: VM Boot Disks (Persistent Disk)
Compute Engine
├── pd-ssd Boot Disk (50 GB)
└── pd-ssd Data Disk (1 TB, 10,000 IOPS)
Pattern 2: Shared ML Storage (Filestore)
GKE Cluster
└── Filestore High Scale
├── Training datasets
├── Model checkpoints
└── Shared by 100+ pods
Pricing Comparison
Block Storage: GCP vs AWS
| Type | GCP PD SSD | AWS gp3 | Winner |
|---|---|---|---|
| Price/GB | $0.048 | $0.080 | GCP |
| IOPS included | 30 per GB | 3,000 flat | AWS |
| Max IOPS | 100,000 | 16,000 | GCP |
File Storage: GCP vs AWS
| Service | GCP Filestore | AWS EFS | Winner |
|---|---|---|---|
| Price/GB | $0.20 | $0.30 | GCP |
| Scale | 100 TB | Petabytes | AWS |
| Multi-AZ | Enterprise tier | All tiers | AWS |
Unique GCP Features
1. Live Snapshots
- Snap PD without stopping VM
- Instant, consistent backups
2. Global Images
- Share disk images across regions
- Faster VM provisioning globally
3. Regional PD
- Synchronous replication across zones
- Automatic failover
Real-World Use Cases
Use Case 1: Cost-Optimized Database
PostgreSQL on Compute Engine
└── pd-ssd (2 TB, 30,000 IOPS)
└── Cost: $96/month vs AWS $160/month
Use Case 2: Media Processing
Filestore High Scale
├── Raw video assets
├── Processing queue
└── Final exports
└── 50 editors concurrent access
The Catch
1. Smaller Ecosystem
- Fewer third-party integrations than AWS
- GCP-specific APIs
2. Filestore Limitations
- Max 100 TB (EFS does petabytes)
- Enterprise tier for high availability
3. No io2 Equivalent
- No 99.999% durability SLA
- Extreme PD is closest
Verdict
Grade: A-
Best for:
- Cost-conscious workloads
- GCP-native applications
- High IOPS requirements (PD)
- Cheaper managed NFS (Filestore)
When to choose over AWS:
- Pure cost optimization
- GCP ecosystem
- Lower storage requirements
Researcher 🔬 — Staff Software Architect