New ITG Report Is Reinforcement For The IBM i Advocate
December 10, 2012 Dan Burger
Choosing one computer system over another has the potential to significantly reduce costs, but each individual company has to do some homework before coming to a decision. It is disappointing, to say the least, that many of these decisions are made regardless of whether they make good business sense. Can a business case for IBM i running on Power Systems be made when compared to Windows and Linux alternatives? Yes, indeed, when bolstered by reports like the one recently released by ITG.
ITG, which is called on regularly to assemble total cost of ownership studies that build a favorable case for IBM‘s midrange and enterprise customer base, has calculated that savings in the range of 40 percent and up are realistic. Total cost of ownership (TCO) adds major indirect costs to the up-front charges for hardware, software (including the operating system and database and other licenses, maintenance for both), and the personnel required to operate the systems.
Much of this TCO data is swept under the rug when organizations make computing decisions, or at least it is downplayed. Some of that is intentional because it doesn’t fit someone’s personal or professional agenda, while some of it is unavailable or assumed to be too difficult to determine.
The total cost of ownership song has been sung by IBM i advocates ever since Unix and X86 machines became a threat in the midrange more than two decades ago, and a similar tune was chanted for System/3X iron in the proprietary minicomputer era that preceded it. Cheaper alternatives at the point of sale necessitated a change of perspective to what most refer to as “the big picture.” It’s a legitimate way to do comparisons. In the automotive market, cheap cars used to be synonymous with poor value. Now that is rarely the case. In computers, particularly enterprise systems, there’s still a case to be made that inexpensive on the front end equates with more expensive over the lifecycle of the equipment.
The latest ITG report makes that case again. I’m not sure how many of these TCO reports ITG has cranked out for IBM, but it’s more than a few. This one goes down essentially the same path as the ones before it. Changes in the market correspondingly bring changes to the report. But the central case that is laid out revolves around the fact that systems running on Windows and Linux require more boxes and more people to manage them and are, therefore, more expensive when the long-term cost of those two factors are taken into account.
The point of the ITG report, whether you want to scrutinize it or not, is that it is a weapon for the unarmed IBM i advocate who is making a business case for this platform. It provides comparisons, examples, and options. And, most importantly it demonstrates a degree of fiscal aptitude and cost structure analysis that can get the attention of the C-level decision makers who do not understand techno geek-speak. Someone who thinks in terms of not only direct but indirect costs for both equipment and labor, as well as how his department can support the organization, and backs up that thinking with a report from an independent authoritative source, makes his or her case stronger.
Advocates for the IBM i platform have a better chance of impressing management with information that represents “the real cost of computing” and that supports the idea investing in a system that reduces cost over a three-year period means the IT budget can afford innovation. Of course, you have to be able to show you have the right stuff to innovate as well.
I’d bet most IBM i advocates already know a fair and balanced comparison between an IBM i system and an Intel system requires more servers and more people to support those servers on the Intel side. But how many can back it up with a report? ITG has done that with Power 720 and 740 systems running IBM i 7.1 with the built-in DB2 for i database and PowerVM hypervisor compared to X86 servers running Microsoft Windows Server 2008, SQL Server 2008, and equipped with the VMware ESXi 5 hypervisor. Other X86 machines match up Linux with the Oracle‘s 11g database and on top of VMware ESXi 5.
The way the systems were configured, as many as ten physical Windows servers, two dedicated as VMware hosts, were used in comparison to the Power Systems 720 and 740 duo.
The configuration was designed to handle production ERP, CRM, e-commerce, business intelligence, and supply chain management applications, plus development and testing. It also featured high availability as the cost of downtime was emphasized in this study.
The Power Systems and X86 server configurations used in the ITG report are represented in this diagram. (Illustration reprinted from the ITG report.) Click graphic to enlarge.
The overall three-year cost averages were pegged at $480,200 for the IBM i setup, $862,200 for the Windows setup, and $1,118,300 for the Linux setup.
Even though the cost of individual X86 servers will always be less expensive than an equivalent Power Systems box, ITG says the number of X86 servers and the necessary software required results in higher acquisition costs. Separate servers were required to handle database, applications, and Web serving, as well as to support test, development, and production instances. Failover clusters were used for Windows database servers and costs on the X86 side were also affected by the price of system administration tools, security, and other functions to provide capabilities equivalent to those included in the base IBM i 7.1 offering.
ITG determined those costs and concluded the initial acquisition costs of the IBM i system was 24 percent less than the Windows system and 47 percent less than the Linux system.
The overall three-year costs per platform give IBM i the advantage in every category, but the greatest savings is attributed to personnel costs. (Illustration reprinted from the ITG report.) Click graphic to enlarge.
The biggest differentiator in the total cost of ownership was attributable to personnel. The cost of managing the Windows setup was 2.3 times higher and the Linux/Oracle setup was 2.6 times higher than the IBM i setup. The difference in the two X86 systems was tied to the higher cost of the Oracle database and the lower productivity of administrators working on the Linux/Oracle system.
There will be people who will find fault with the way this report was set up and the way the comparisons were made. It is a fair criticism based on the many reports that are commissioned and conducted on behalf of IT companies that carefully plan what the outcomes should be.
“The purpose of the report is to provide general guidelines,” Ian Jarman, analytics marketing manager for Power Systems and System z and a former IBM i product manager, told me last week on the phone. “Does the IBM client have more or less servers [than these competitors]? There are always fewer IBM i servers. This report puts structure around that. It identifies the impact in other areas such as maintenance, software licenses, support, personnel, and facilities.”
Jarman says these are the key costs and should be the focus of discussions about investing in business applications and how to manage IT efficiently. He defends the ITG report calling it well researched and expressed confidence in the review and the data that goes into it.
“ITG has a strong reputation and the skills and experience to do research across multiple platforms,” Jarman said. “We give a framework to ITG for research, and we review their research. We make sure they portray IBM i as we view it, but it is researched independently.”
Allison Butterill, IBM i product offering manager, was on the same conference call with Jarman and myself. She pointed out that it’s the IBM i shops that request these reports are researched and published.
“Groups like the COMMON advisory council and the LUG (large user group) use them to justify [to upper-level management] new projects or upgrades. These reports help them make business cases to management. They are valuable and they focus on different priorities. Our customers are adamant that this is important for them.”
The entire report can be seen here.