High-performance block storage for VMs with Hyperdisk and Persistent Disk for flexible performance scaling.
What is Google Cloud Block Storage?
Google Cloud Block Storage covers the network-based block storage options for Compute Engine VMs. Google distinguishes between Hyperdisk, the recommended and highest-performing option, and Persistent Disk, maintained as a legacy offering with performance proportional to provisioned capacity. Persistent Disk is no longer available on the newest machine series; Google recommends Hyperdisk for new deployments.
Hyperdisk decouples performance from capacity: IOPS and throughput can be configured independently. The product family includes Hyperdisk Balanced for most workloads, Hyperdisk Extreme for applications with very high IOPS requirements, Hyperdisk ML for fast loading of large datasets during machine learning training, and Hyperdisk Throughput for sequential, throughput-oriented workloads.
Both Hyperdisk and Persistent Disk offer encryption at rest and in transit, can be managed independently of individual VMs, can be moved between instances, and achieve high availability through cross-zone replication.
Core Features
- Flexible Performance: Hyperdisk enables independent configuration of IOPS and throughput, decoupled from capacity
- High Availability: Cross-zone replication and portability between VM instances without downtime
- Encryption: Encryption at rest and in transit, optionally with Customer-Managed Keys (CMEK)
- Independent Management: Volumes can be managed independently of the lifecycle of individual compute instances
Typical Use Cases
Database Workloads with Variable Requirements
Databases with changing load profiles benefit from Hyperdisk Balanced or Extreme. IOPS and throughput can be adjusted independently of capacity, enabling on-demand performance with predictable costs.
ML Training with High Throughput
Machine learning training often requires high sequential throughput for loading large datasets. Hyperdisk ML is specifically designed for these workloads and, per Google’s documentation, supports significantly higher per-volume throughput than classic Persistent Disk offerings.
Benefits
- Independent scaling of performance and capacity with Hyperdisk
- A range of storage types for different workload profiles
- Seamless integration with Compute Engine and Google Kubernetes Engine
- High availability through cross-zone replication
Integration with innFactory
As a certified Google Cloud partner, innFactory supports you with Block Storage: storage architecture design, migration from existing block storage solutions, selecting the right Hyperdisk type, and performance and cost optimization.
Typical Use Cases
Frequently Asked Questions
What is Google Cloud Block Storage?
Google Cloud Block Storage refers to the network-based block storage options for Compute Engine VMs. Google now recommends Hyperdisk for most workloads, offering independently configurable performance; Persistent Disk is maintained as a legacy option whose performance is tied to provisioned capacity and is no longer available on the newest machine series.
What is the difference between Persistent Disk and Hyperdisk?
Persistent Disk scales performance with disk size. Hyperdisk decouples performance from capacity, allowing independent configuration of IOPS and throughput plus higher performance limits. Hyperdisk is recommended for new projects and current machine series.
What Hyperdisk types are available?
Hyperdisk Balanced for most workloads with a strong price-performance balance, Hyperdisk Extreme for workloads with very high IOPS requirements, Hyperdisk ML for loading large datasets during ML training with very high throughput, and Hyperdisk Throughput for sequential, throughput-oriented workloads.
How is Google Cloud Block Storage billed?
Billing is based on provisioned storage capacity from creation until the volume is deleted. For Hyperdisk (aside from included Throughput baselines) and for Persistent Disk Extreme, provisioned IOPS and throughput are billed separately. Storage is billed even when a volume is attached to a stopped instance or unattached.
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of Google Cloud (official documentation). This page does not represent an offer by Google Cloud.
