Hybrid NVMe + GCS architecture with Active-Active HA delivers up to 11.56 GB/s NFS client throughput. Choose from validated configurations optimized for your workload. Deploy in minutes with Terraform.
One-click deployment from GCP Marketplace or instant Terraform provisioning. No complex configuration or storage expertise required.
Built on Google Cloud Storage (GCS) for truly unlimited capacity. Scale from gigabytes to petabytes without pre-provisioning.
Up to 11.56 GB/s NFS client read throughput with Active-Active HA. ZFS local storage delivers 2.0-4.7 GB/s per-node. Choose from validated configurations optimized for your workload.
ZFS-based architecture with snapshots, compression, deduplication, and replication. Active-Passive HA with sub-60 second failover.
Runs on GCE instances (N2, N2D families) with NVMe persistent disks for optimal performance.
Leverages GCS buckets for unlimited bulk data storage with cost-effective pricing tiers.
Fully integrated with VPC networking, Private Google Access, and Cloud Load Balancing.
Mount as persistent volumes for GKE workloads with standard NFS protocol support.
Ideal for ML training data lakes with high-throughput access to large datasets.
Full integration with Cloud Logging, Cloud Monitoring, and alerting capabilities.
MayaNAS uses a hybrid architecture that combines the best of both worlds:
Hybrid architecture with NVMe persistent disks for metadata and Google Cloud Storage for unlimited capacity. Both nodes actively serve NFS traffic for maximum throughput.
Store massive training datasets with high-throughput access for Vertex AI, GPU instances, and ML pipelines. Unlimited capacity scales with your data growth.
Cloud-based video editing, rendering, and transcoding with Cloud GPUs. Handle 4K/8K media files with consistent performance.
Shared home directories, Git repositories, build artifacts, and development tools for distributed teams on GCE or GKE.
Staging area for BigQuery data loads, analytics workflows, and large-scale data processing with Dataflow or Dataproc.
One-click deployment directly from GCP Marketplace.
Infrastructure-as-Code for automation and GitOps workflows.
All performance numbers are validated from actual benchmark tests on Google Cloud Platform. Choose the right configuration for your workload - all use hybrid NVMe + GCS architecture with unlimited capacity.
3.1 GB/s Write | 4.7 GB/s Read | n2-standard-48 (48 vCPUs, 192GB) + 10 GCS Buckets | ~$1,650/mo | AI/ML training, maximum throughput workloads
2.9 GB/s Write | 3.5 GB/s Read | n2-standard-32 (32 vCPUs, 128GB) + 10 GCS Buckets | ~$1,100/mo | Large datasets, media production
2.3 GB/s Write | 2.9 GB/s Read | n2-standard-16 (16 vCPUs, 64GB) + 6 GCS Buckets | ~$550/mo | Production workloads, shared storage
2.0 GB/s Write | 2.3 GB/s Read | n2-standard-16 (16 vCPUs, 64GB) + 6 GCS Buckets | ~$550/mo | Development, testing, small teams
| Configuration | Instance | vCPUs | RAM | GCS Buckets | Write (GB/s) | Read (GB/s) | Est. Cost | Use Case |
|---|---|---|---|---|---|---|---|---|
| Entry | n2-standard-16 | 16 | 64 GB | 6 | 2.0 | 2.3 | ~$550/mo | Development, testing, small teams |
| Standard | n2-standard-16 | 16 | 64 GB | 6 | 2.3 | 2.9 | ~$550/mo | Production workloads, shared storage |
| High | n2-standard-32 | 32 | 128 GB | 10 | 2.9 | 3.5 | ~$1,100/mo | Large datasets, media production |
| Ultra | n2-standard-48 | 48 | 192 GB | 10 | 3.1 | 4.7 | ~$1,650/mo | AI/ML training, maximum throughput |
Cost estimates include compute instance, NVMe persistent disk for metadata, and GCS storage. Actual costs may vary based on data volume and region. Tested with 10G network, fio benchmarks (4K random, 1M sequential).
End-to-end NFS client throughput with Active-Active HA - validated performance from actual deployments on Google Cloud Platform
| Single Node Read: | ~5.5 GB/s |
| Single Node Write: | ~1.4-2.2 GB/s |
| Combined HA Read: | 11.56 GB/s |
| Combined HA Write: | 4.72 GB/s |
Configuration: n2-standard-32 (32 vCPUs, 128GB) + 24 GCS buckets, Active-Active HA
| Single Node Read: | ~4.0 GB/s |
| Single Node Write: | ~2.0 GB/s |
| Combined HA Read: | 8.14 GB/s |
| Combined HA Write: | 4.18 GB/s |
Configuration: n2-standard-48 (48 vCPUs, 192GB) + 20 GCS buckets, Active-Active HA
With Active-Active HA deployment, both nodes serve NFS traffic simultaneously, effectively doubling throughput compared to single-node. The graphs above show individual share performance (S1/S2) and concurrent combined performance, demonstrating how MayaNAS scales with Active-Active architecture.
Point-in-time snapshots with zero performance impact. Perfect for backup and recovery workflows.
Transparent compression reduces storage costs by 30-50% depending on data type.
Cross-region replication for disaster recovery and multi-region deployments.
ZFS checksums detect and correct silent data corruption automatically.
NFSv3, NFSv4 support for Linux, macOS, and Windows clients.
Integration with Cloud Monitoring for performance metrics and alerting.
Validated 11.56 GB/s NFS client performance with Active-Active HA—powered by Google Cloud Storage for cloud-native economics. Combine enterprise-grade performance with cost-efficient object storage tiers. Choose from proven configurations optimized for your workload and deploy from GCP Marketplace or contact our team for custom configurations.
Get Started Contact SalesMayaNAS configurations range from $550-$1,650/month depending on instance size and GCS bucket count. Contact our sales team for detailed pricing and custom configurations. Available on GCP Marketplace with billing through your GCP account.
More GCS buckets increase parallel throughput. Entry/Standard use 6 buckets (2.0-2.9 GB/s), while High/Ultra use 10 buckets (2.9-4.7 GB/s). Bucket count is configured during deployment.
ZFS local storage performance (2.0-4.7 GB/s shown in carousel) measures storage backend throughput per node. NFS client performance (up to 11.56 GB/s) measures actual throughput experienced by NFS clients. With Active-Active HA, both nodes serve traffic simultaneously, effectively doubling client-visible performance.
Yes, you can resize instances or redeploy with different configurations. Data is stored in GCS buckets and persists across redeployments.
MayaNAS can be deployed in any GCP region that supports Compute Engine N2/N2D instances and Google Cloud Storage.
Data in GCS is automatically replicated. For additional protection, use ZFS snapshots and cross-region replication features.
MayaNAS runs on standard GCP services: Compute Engine instances (N2/N2D), NVMe persistent disks, and Google Cloud Storage. Standard GCP infrastructure costs apply.
Yes, contact our sales team to discuss trial options and proof-of-concept deployments tailored to your specific use case.