Hyperconverged Infrastructure: An Overview

Introduction

• Hyperconverged Infrastructure (HCI) promises reduced costs, greater flexibility.
• It's essential to compare it with traditional datacentre infrastructure to understand its benefits.

Traditional Datacentre Infrastructure

• Also known as a three-tier infrastructure:
  1. Network Tier
  2. Server Tier
  3. Storage Tier

Standalone Server

• Combines computation (processors and memory) and storage (hard drives) in one package.
• Inefficient for scaling: managing multiple units, over-provisioning storage.

Storage Area Network (SAN)

• Separates storage from computation resources.
• Servers do computation; connect to centrally managed storage pool.
• Benefits:
  • Increased flexibility and scalability.
  • Easier replication and backup.

Virtualization

• Decouples computation from hardware.
• Logical servers can move between physical host servers.
• Requires separate storage tier for mobility.

Traditional Three-Tier Infrastructure

• Network, server, and storage tiers operate independently.
• Managed by different teams.

Converged Infrastructure

• Packaged product by a single vendor (e.g., Cisco's FlexPod).
• Bundles network, server, and storage.
• Single vendor support and compatibility.
• Downsides:
  • Vendor lock-in.
  • Potentially expensive scalability.

Hyperconverged Infrastructure

• Collapses server and storage tiers into one.
• Uses commodity x86 hardware interconnected via IP network switches.
• Virtualization allows logical servers to move between physical hosts.
• Data is distributed like a SAN.
• Benefits:
  • Reduced hardware costs.
  • Simplified scaling.
  • Simplified management.

Administration

• Varies by vendor; some offer single management interfaces, others are more DIY.
• Potentially reduced administrative overheads.

Potential Downsides of HCI

• Scalability Issues:
  • Not all applications scale linearly.
  • Example: File servers increase storage needs but not computation.
• Complexity:
  • Combined computation and storage complexities.
  • Maintenance challenges (e.g., taking down a virtualisation host affects storage).

Considerations

• Depends on infrastructure's scaling nature.
• Linear scaling could save hardware costs; non-linear could increase costs.
• Mixed approach might be suitable for different environments (e.g., main datacenter vs. branch site).
• Performance characteristics of workloads are crucial.

Conclusion

• No one-size-fits-all answer.
• Must evaluate based on specific needs and scenarios.

Call to Action

• Leave comments on experiences with hyperconverged infrastructure.
• Like and subscribe for more content.

Thanks for watching!