Hybrid HCI-Cloud: Extending to Public Cloud
A retail chain with 12 branches in the Philippines had a problem: their Nutanix cluster handled daily operations perfectly, but during quarterly inventory uploads, all 12 branches flooded the system simultaneously. They needed burst capacity for 48 hours every quarter.
Buying extra Nutanix nodes for 48 hours of peak usage per year made no financial sense. Instead, we extended their HCI to AWS — burst compute capacity on-demand, pay only for what they use. Their quarterly peaks went from 4-hour processing windows to 45 minutes.
That's the hybrid HCI-cloud story: keep your core workloads on-premises, burst to the cloud when you need it.
What is Hybrid HCI-Cloud?
Hybrid HCI-cloud connects your on-premises HCI cluster to public cloud resources (AWS, Azure, or GCP). It's not a full cloud migration — it's an extension. Your core workloads stay on-premises where they're fast and cheap. You use cloud resources for specific use cases.
Think of it like this: your HCI is your home kitchen (fast, familiar, always available). The cloud is a catering service (great for big events, but you wouldn't use it for daily cooking).
Use Cases for Hybrid HCI-Cloud
Here are the four most common hybrid patterns we deploy:
1. Disaster Recovery (DR) to Cloud:
Replicate critical VMs from on-premises HCI to cloud (AWS Outposts, Azure Stack HCI, or native EC2/VM).
RPO: 15-60 minutes. RTO: 30-60 minutes.
Cost: Pay only for DR compute when it's running (cloud DR is 60-80% cheaper than a physical DR site).
2. Cloud Bursting:
Use cloud compute for peak workloads (quarterly processing, seasonal traffic, batch jobs).
On-demand scaling: add 10-50 VMs in the cloud for 48 hours, then shut them down.
Perfect for retail (holiday peaks), finance (quarterly closes), and manufacturing (batch processing).
3. Development/Test in Cloud:
Spin up dev/test environments in the cloud instead of provisioning on-premises resources.
Developers get self-service provisioning. No waiting for IT.
Cost: dev/test in the cloud costs 50-70% less than dedicated on-premises hardware.
4. Cloud-Native Integration:
Connect on-premises HCI to cloud-native services (AWS S3, Azure AI, GCP BigQuery).
Keep data on-premises, use cloud services for processing.
Best for: analytics, AI/ML inference, SaaS integrations.
Architecture Design
Here's the architecture we use for a typical hybrid HCI-cloud deployment:
On-premises (Primary):
Nutanix/Sangfor HCI cluster running production workloads.
Nutanix Xi or third-party DR solution for cloud replication.
Dedicated 100Mbps-1Gbps link to cloud (AWS Direct Connect or Azure ExpressRoute).
Cloud (Secondary):
AWS VPC or Azure VNet with pre-configured networking.
Warm standby instances for DR (t3.medium or B2s — cheap, ready to scale).
Cloud storage (S3 or Blob) for backup and archival.
Networking:
Site-to-site VPN as backup link.
Direct Connect/ExpressRoute for primary link (lower latency, higher bandwidth).
DNS-based failover for automatic switchover.
Cost Optimization
Hybrid cloud costs can spiral if not managed. Here's how to keep costs down:
Use reserved instances for predictable workloads. If you know you need 10 VMs running in the cloud 24/7, reserved instances save 40-60% versus on-demand.
Use spot instances for batch processing. If your workload can tolerate interruptions (data processing, CI/CD), spot instances cost 70-90% less.
Shut down DR instances when not testing. Cloud DR in warm standby costs $200-500/month. Running it 24/7 for "just in case" is wasteful.
Monitor cloud spending. Use AWS Cost Explorer or Azure Cost Management. Set billing alerts. Review monthly.
A typical hybrid HCI-cloud deployment costs $500-2,000/month in cloud fees, depending on the use case. Compare that to $50,000+ for a physical DR site.
Migration Considerations
Moving workloads between on-premises and cloud requires planning:
Bandwidth: A 100Mbps link takes ~2.5 hours to transfer 1TB. A 1Gbps link takes ~15 minutes. Size your link for your migration timeline.
Licensing: Some software licenses don't cover cloud deployment. Check with your vendors.
Data sovereignty: Some data must stay on-premises (regulations). Know your compliance requirements.
Application dependencies: Don't move a database without its application servers. Map dependencies first.
Best Practices
Start with DR-to-cloud. It's the lowest-risk hybrid pattern. Test failover before relying on it.
Use infrastructure-as-code (Terraform, CloudFormation). Cloud resources should be reproducible, not manually configured.
Monitor everything. Cloud costs, replication health, network performance — dashboards for all of it.
Document the runbook. When you need to failover to cloud at 2am, you want step-by-step instructions, not tribal knowledge.
Review monthly. Cloud costs change. Usage patterns change. Review and optimize every month.
Conclusion
Hybrid HCI-cloud is not about replacing on-premises infrastructure. It's about extending it intelligently. Use cloud for DR, burst capacity, dev/test, and cloud-native services. Keep your core workloads on-premises where they're fast and cost-effective.
Start with DR-to-cloud. It's the easiest entry point and provides immediate value. Once you're comfortable, explore cloud bursting and dev/test use cases.
Want to go deeper? Explore [VMware alternatives](/en/vmware-alternative), [Run infrastructure services](/en/products/run), or [platform comparison](/en/compare).
FAQ
Q: What internet speed do I need for hybrid HCI-cloud?
A: Minimum 100Mbps dedicated link for DR replication. 1Gbps or higher for cloud bursting and large data transfers. AWS Direct Connect or Azure ExpressRoute recommended.
Q: Which cloud is best for hybrid HCI?
A: Depends on your use case. AWS has the most services. Azure integrates best with Microsoft workloads. GCP excels at data analytics. All three work well with Nutanix and VMware HCI.
Q: Can I run Nutanix in the cloud?
A: Yes. Nutanix offers NC2 (Nutanix Cloud Clusters) on AWS and Azure. It extends your Nutanix management plane to the cloud. But it requires significant cloud investment.
Q: How much does hybrid cloud cost monthly?
A: For DR-only: $200-500/month. For DR + burst capacity: $500-2,000/month. For full hybrid: $2,000-10,000/month. Depends heavily on workload and cloud provider.
Sizing and Capacity Planning
Proper sizing is critical for HCI deployments. Start by inventorying your current workloads: CPU cores, memory per VM, storage per VM, and IOPS requirements. A general rule of thumb: each HCI node should run at 60-70% capacity to allow for growth and failover.
For a typical deployment of 50-100 VMs, we recommend starting with 4 nodes, each with: 2x 16-core CPUs, 256GB RAM, 4x 1.92TB NVMe SSDs, and 2x 25GbE NICs. This provides enough resources for most small-to-medium workloads with room to grow.
Migration Strategy from Traditional Infrastructure
Migrating from traditional SAN/NAS-based infrastructure to HCI requires careful planning. We recommend the following approach: First, identify non-critical workloads for initial migration (development, testing, staging environments). Second, use live migration tools (HCX for VMware, Xi Frame for Nutanix) to move VMs with zero downtime.
Third, validate performance on HCI before migrating production workloads. Monitor for 2-4 weeks to ensure IOPS, latency, and throughput meet requirements. Fourth, migrate production workloads in phases, starting with the least critical and progressing to mission-critical systems.
Disaster Recovery with HCI
HCI provides built-in high availability within a cluster, but you still need a disaster recovery plan for site-level failures. Options include: HCI-to-HCI replication between data centers (RPO as low as 5 minutes), cloud-based DR using HCI vendor cloud services, and hybrid DR with cloud object storage for backup.
We typically recommend a 3-2-1 backup strategy: 3 copies of data, on 2 different media types, with 1 copy offsite. With HCI, this translates to: local vSAN replication (copy 1), backup to secondary storage (copy 2), and cloud backup (copy 3).