DIY Storage vs. Hyper-converged Systems

Table of Contents

  1. Summary
  2. What Needs to Be Fixed
  3. Hardware Selection for Software-Defined Storage
  4. DIY Eliminates Storage Problems and Costs
  5. Hyper-convergence: Simplicity Delivered
  6. Eliminating Storage Problems with Hyper-convergence
  7. DIY vs. Hyper-convergence
  8. Summary and Key Takeaways
  9. About Marc Staimer


The hyperscale infrastructure used by Amazon, Google, Microsoft, and Facebook delivers extreme performance that upends the way service providers and IT managers do business. Using commodity servers, networking, storage, and a DIY approach, enterprise IT managers are able to respond to the demands for mobile and cloud.

Historically, the IT pendulum swings between divergence (using best-of-breed components) and convergence (turnkey completely integrated and engineered systems) every 5 to 10 years. DIY is a swing back to divergence, as are the software-defined data center, elastic infrastructure, bare-metal infrastructure, containers, and object storage.

What makes the current trend unique is that while divergence is gaining momentum, so is convergence with hyper-convergence, engineered systems, appliances, and software-defined compute/networking/ storage that is shrink-wrapped in hardware. Most IT managers see the paths as contradictory, but that is not always the case.

This report explains which approach offers the best short-term and long-term results in operational efficiencies, adaptability, manageability, and the total cost of ownership (TCO).

Key highlights from this report include:

  • Financially unsustainable, storage costs are now the largest single-line item in the IT budget.
  • DIY storage minimizes costs significantly by using off-the-shelf server components.
  • Hyper-convergence improves storage scalability and manageability, but introduces potential vendor lock-in.
  • No one answer is right for every IT organization or every storage project.

Thumbnail image courtesy of camij/iStock.

Access Report

Available to GigaOm Research Subscribers

Subscribe to
GigaOm Research