Top IBM Power Systems myths: “Power Systems are too expensive compared to x86”

28 May, 2019
Skip Garvin
IBM

There are many misconceptions about IBM Power Systems in the marketplace today, and this blog series is helping to dispel some of the top myths. My last post put aside the myth that the OpenPOWER Foundation isn’t really an industry backed consortium and has no real influence or value in the marketplace.

In this post, I’d like to address the myth that Power Systems are too expensive when compared to x86 systems. In exploring this myth, I’ll focus on on-premises solutions. Cloud pricing is very different, and a topic that warrants its own article.

Reality and marketplace perceptions

Unlike other myths we’ve explored in this series, there’s an element of truth in this one. There are scenarios where the cost to acquire IBM Power Systems servers will be higher than the cost to acquire an x86 system. Conversely, there are scenarios where the cost to acquire a Power-based solution will be less.

There’s a general perception that IBM Power Systems are not competitively priced when compared to x86, but this is based on an oversimplified understanding of how server pricing works. In fact, all organizations have their own process for determining a competitive price based on their priorities. The criteria may include:

  • Application requirements
  • Total cost of acquisition
  • Total cost of ownership
  • Advice from industry consultants
  • Special bids

What matters more than cost

Industry studies indicate that lowest cost may not be the number one criterion for organizations acquiring servers today. Consider the recent IDC Enterprise Infrastructure Market Pulse 1Q 2019 Vendor View study, which asked clients to identify the top criteria for selecting servers in their organization. Surprisingly, 32 percent of respondents ranked “Technology Innovation” as a key criterion. Here are a few other rankings from the study:

  • Reliability: 48 percent
  • Performance: 42 percent
  • Scalability: 28 percent
  • Capital Cost: 25 percent
  • Total Cost of Ownership: 24 percent

You can see here that cost is often not the top priority.

This IDC report supports an internal study IBM did in May 2018 that showed that only 3 percent of clients surveyed ranked “lower cost” as the most important reason for selecting servers.

Regardless of where you rank the importance of lower cost in your selection criteria, I hope this post will help you understand that IBM Power Systems offers very competitively priced servers for the applications driving innovation in the marketplace today.

Consider this quote from the CIO of a very large company IBM worked with during a migration project:

 “The most important element of any system I buy is the continuous, uninterrupted operation of my applications. I don’t like surprises.”

It clearly articulates his priority of reliable performance, and it’s not dissimilar to what we saw in the IDC study and the IBM internal study. Better performance, more reliability, more security and better scalability often rank ahead of lower cost as criteria for acquiring servers. Having a competitive price that provides value to clients is essential to our success — but that’s very different than just having the lowest price.

Total cost of acquisition (TCA) or total cost of ownership (TCO): Which one matters?

No one pays list price for these systems, and the discount structures for each client are different. Special bids are commonplace. There are thousands of “point in time” slides that do price comparisons using different benchmarks, application suites and configurations, but these become dated very quickly. What do most organizations use?

In 2018, Solitaire Interglobal did a study titled “SAP HANA on Power” comparing it to similar x86 systems. Here are a few points from the study to consider:

  • Many companies only focus on the initial acquisition cost of a system (TCA) rather than the overall cost of ownership (TCO).
  • The TCA used by many of these companies was the initial quote from the vendor.
  • Initial vendor quotes varied from the actual implementation cost by as much as 35 percent.
  • A large majority of the IBM SAP HANA quotations showed smaller variations between the initial sizing and the actual deployment than other vendors — about 2.8 percent.

Companies spent considerably more than the initial quote from the vendor to get their x86 system up and running. These findings are not unique. Experience has shown us that companies often have to spend more to get an x86 system running than they would for Power Systems.

TCO has always been one of the leading business metrics for operational efficiency because it measures ongoing production costs. The Solitaire Interglobal study found that the overall cost of ownership for IBM Power-based implementations is lower by as much as 78 percent compared to x86 systems — and this finding is also not unique to this study.

The IDC study referenced above also reinforces the importance total cost of ownership versus just the acquisition cost when buying servers for your organization.

How Power Systems offer a better TCO

Here are a few reasons why IBM Power Systems offer a lower TCO to clients:

  • Higher per-core performance means fewer cores to run the same workloads.
  • Fewer cores means lower software licensing costs.
  • Smaller physical footprint (fewer systems) means reduced power and cooling costs.
  • Fewer servers to manage means lower system management costs.
  • Resilience – the ability of the platform to absorb unplanned workload on a temporary or short-term basis without overall platform failure.

Low acquisition costs feel good initially, but investing in a solution that’s going to cost more to operate over the life of that system is never a good business decision.

Kevin Rakos, Systems Administrator at Earth Signal, offered the following observation:

“I think that the initial cost to acquire the units may be a little more, but the benefits outweigh that upfront cost. In our industry, other companies have gone down the route of cobbling together as many x86 processors as they can, and the result is a sprawling, energy-hungry infrastructure. The IBM POWER architecture gives us far more performance than x86 in a much more compact and efficient form-factor.”

See the full Case Study for details.

Ready to learn more?

“Have you looked at Power lately?” is a question I’d ask anyone tasked with acquiring systems today. There are far too many combinations of system configurations and workloads to present any meaningful price comparisons in this post, so please take some time to contact us.

We can provide detailed comparisons, so you can see just how competitively priced our Power Linux, AIX and IBM i series systems are when compared to an x86 equivalent. (To learn more about the value IBM i has been providing our clients for decades, enjoy this story.)

IBM Systems Lab Services has a team of experienced consultants ready to help you get the most out of IBM Power Systems. Reach out to us today if you have questions and stay tuned for the next post in this series.

The post Top IBM Power Systems myths: “Power Systems are too expensive compared to x86” appeared first on IBM IT Infrastructure Blog.