Executive Summary
There are three primary objections to mainframe ownership: price, skills and “old technology.” Of these, the pricing objection has been the most difficult to overcome. (The skills issue is being addressed through academic initiatives and software simplification; the “old technology” argument is simply not true – mainframe technology is highly advanced compared to most “modern” x86 server designs.)
For years the basic tenant in IBM’s Z mainframe pricing has been based upon charging customers for sub-capacity usage, using a metric that tracks the highest “rolling 4 hour average” (R4HA) in any given billing period. Mainframe customers are, accordingly, billed on the system capacity used at the highest usage rate during a given billing cycle. This practice has forced IBM mainframe customers to manically focus on balancing the capacity use of their mainframe in order to avoid R4HA computing peaks that drive up mainframe compute costs.
Meanwhile, R4HA pricing has served to inhibit mainframe market growth – especially in the hybrid cloud marketplace. Fearing high utilization charges, IT managers have regularly moved applications off their mainframes to other platforms in order to make room for existing mission-critical applications. Instead of encouraging workload growth and expansion on the mainframe, R4HA has had the opposite effect, driving workloads, applications and development environments to other platforms.
To address this discontinuity, IBM has announced “Tailored Fit Pricing for IBM Z,” a new pricing model that comes in two flavors: 1) an Enterprise Consumption Solution; and, 2) an Enterprise Capacity Solution. The former is priced on the total number of MSUs (millions of service units) consumed – no longer the number used at system peak. The latter is based on a commitment to use a fixed amount of capacity – and includes greatly discounted pricing for new workloads.
The Enterprise Consumption Solution is geared for mainframe users with variable workloads – and allows them to pay for the computing power they use in total, not at maximum peak. The Enterprise Capacity Solution is geared for mainframe users that prefer predictable and consistent monthly charges – with no surprises due to higher-than-expected peak utilization.
From our perspective at Clabby Analytics, these new pricing models are designed to simplify mainframe systems and budget management while also driving mainframe expansion and growth. Each allows customers to consume mainframe computing power in a manner that makes sense to their business imperatives. And, we expect that IBM incentives for placing new workloads will serve to make mainframes more competitive with other cloud offerings, positioning them for solid growth in hybrid cloud environments. These new models are a “win/win” for IBM customers, and for the maker of the venerable mainframe: IBM.
IBM’s R4HA – Ready for retirement
We’ll not sugarcoat the following paragraph. IBM’s R4HA pricing model has long been horrendous. It has been a model that has forced mainframe users to focus to the extreme on managing workloads to minimize capacity usage – and activity that does not drive a business forward – but rather adds additional tasks to an already overburdened mainframe management workforce. It puts an extreme and unnecessary focus on machine tuning and monitoring. Goodbye – and good riddance – to R4HA pricing!
The good news is that IBM has known for a while that R4HA pricing has been a universal pain point for its mainframe base – and has been working hard to address the problem. Over the past few years, Clabby Analytics has seen IBM experiment with several different mainframe pricing models, including, most notably, Container Pricing. This was a step-in-the-right-direction – providing the transparency, predictability, and flexibility that IBM customers have been seeking.
Container pricing was the forerunner to Tailored Fit Pricing. Unlike in the Docker world, IBM’s Container Pricing did not actually use the Linux container technology. Rather it was simply a mechanism IBM designed to track specific workloads running on a mainframe – enabling the company to charge different prices for differing workloads. With container pricing, IBM was able to lower the price to run certain “approved” workloads, most notably development and test workloads, as well as new z/OS-based applications. This gave the mainframe the ability to run workloads at the same price (or even lower) as x86-based or public cloud competitors. Other workloads, however, remained tied to R4HA pricing.
In 2018, IBM enhanced Container Pricing with Solution Consumption License Charges (SCLC) , which offered mainframe customer the ability to “pay-as-you-go,” with a usage-based pricing model aimed at helping the mainframe acquire new workloads. But again, for existing workloads, R4HA peak-load capacity remained the dominant pricing scheme. Accordingly, neither pricing scheme truly addressed the mainframe client’s major pain point – peak load capacity pricing.
IBM’s new offerings
Mainframe customers can remain with R4HA pricing if they wish – but next month, IBM will formally roll-out its Enterprise Consumption and Enterprise Capacity solutions (general availability on June 21st) as comprehensive alternatives to the R4HA metric it has used to charge its customers for the last 20 years or so. These two solutions are being launched as the Tailored Fit Pricing generation of IBM Z software licensing. The aforementioned container pricing solutions are also being rebranded under the Tailored Fit Pricing name – finally ending the confusion with other types of ‘containers’ in the IT world.
For customers who chose the Enterprise Consumption pricing model, IBM will no longer be using the E5/B5 sections of the SCRT reports to charge based on R4HA peak usage data. Instead, there will be a new Tailored Fit Pricing SCRT report that will charge based on total MSUs consumed. Mainframe administrators can see what their current MSUs consumed are in the N7 section of their SCRT report today. Customers who have spikey, variable workloads – such as retailers during sales events, or health insurance providers during enrollment periods – will simply pay for a mutually-agreed-to amount of MSU computing power used annually.
Plus, to sweeten this deal, at the end of a given year, if a mainframe customer uses more computing power than projected, that over-usage is considered new growth – and is eligible for growth pricing that is typically 50% of the effective per-MSU price. Further, if a customer uses fewer MSUs than projected, that excess will be credited to the following year. Congratulations, IBM, you’ve finally found a way to incent existing customers to ADD new workloads to their mainframe base!
Customers who chose the Enterprise Capacity pricing model (typically customers who have steady, predictable workloads such as banks and government institutions) agree to use a fixed amount of capacity billed at simple, predictable monthly rates. These customers must also make a commitment to deploy a new workload or expand an existing workload – but, as is the case with the Enterprise Consumption model, that increase in utilization is also steeply discounted with aggressive growth pricing. These customers also gain additional flexibility because they are able to reconfigure their technical environments at will without price changes. And they are incented with allowances for increased development and test activities. Kudos, again, to IBM!
The bottom line
IBM’s mainframe organization has long known that its R4HA pricing metric has been unpopular. And it tried several different approaches to simplify mainframe pricing for its customers – while also encouraging the deployment of new applications on mainframes. These new capacity and consumption approaches represent a logical approach for serving two different classes of mainframe users – those with variable workloads and those with steady workloads.
In essence, these new approaches are in line with the way that IBM mainframe customers use their mainframes – and in line with the way they prefer to pay for mainframe usage. But what we like best about these new pricing schemes is IBM’s willingness to greatly discount not just new workloads, but also growth of existing workloads. As we said at the outset, the R4HA approach focused mainframe users on capacity management – and resulted in the movement of applications off of the mainframe to other architectures in the cloud in order to avoid steep peak usage charges. These new models encourage exactly the opposite – all workload growth on the mainframe platform.
We also like the effect that the new pricing schemes will potentially have on mainframe management. There are a raft of tools and utilities that IBM mainframe managers have been forced to use to spend their time managing capacity – a whole industry of software and educational services has grown-up around mainframe capacity management. Now, capacity will still have to be measured – but gone is the price penalty for running a mainframe at peak capacity – and that’s a real big deal. Mainframe managers can spend less time fine tuning mainframe capacity use, and spend more time on other business-advancing activities (such as deploying new workloads; strengthening security, and so on.)
Summary observations
For years, IBM mainframe customers have been asking for less complex pricing and additional pricing flexibility when deploying new workloads. They did not like having to jump through hoops with requests for special pricing adjustments for new workloads– and they did not take to specialized configurations. This new approach gives mainframe users something they have long wanted: a simple, logical approach to mainframe compute pricing.
This new model is also a big win for IBM. It gives the company’s Z sales force a means to encourage new application deployment on mainframes. Usage will grow, and more overall capacity will be sold (IBM’s real goal.) Finally, it should be noted that IBM has beta-tested its consumption solutions with sixteen customers to date – and nine customers with capacity solutions. Expect these numbers to grow rapidly as mainframe users rush to adopt IBM’s new “Tailored Fit” mainframe pricing models.