Workload patterns are constantly changing in the age of hybrid and multi-cloud. To survive, every vendor on the planet must listen to customer needs and respond. One thing that clients and the industry all agree on is the future is very much hybrid cloud spanning on-prem to the public cloud. That is the new normal.
I have been covering the progress of the IBM Z and positioning it as a new class of computing platform that is open and well suited to drive the digitally transformed business. I would encourage you to dispel any preconceived notions of mainframe technology because the IBM Z is constantly evolving to support its enterprise customers' needs and business requirements. The speed of the adoption of cloud technologies by Z has been fast and I dare say… not like most things used to be in the mainframe world.
An area where IBM wasn’t at the front of the line was how it charged enterprises for using mainframe software and hardware. This shortcoming has now been changed with the announcement of flexibility around hardware pricing to complement the new pricing model for software called Tailored Fit Pricing for IBM Z software announced in May 2019.
The acronym soup of pricing and why it needed to change
For almost twenty years, IBM used the million service units (MSU) metric for software pricing purposes. MSU is a measure of the amount of CPU consumed per hour. Monthly license charges (MLC) used the peak rolling four-hour average (R4HA) as reported by the Sub-Capacity Reporting Tool (SCRT), which worked fine back in the 1990s when workloads were reasonably predictable.
As workloads became related to commercial opportunities with usage spikes, this pricing model drove IT to manage usage to avoid the peak R4HA consumption and charges, a practice known as capping.
Something needed to change as the pricing model was driving the wrong behavior, stifling innovation, discouraging the deployment of new workloads, and limiting the ability for businesses to leverage the total value of Z technology, not to mention the unpredictable monthly billing model.
A new compute-as-a-service pricing model for the hybrid cloud
The IBM Z pricing model introduced in May 2019 named Tailored Fit Pricing (TFP), which now includes a where the highly successful Software Consumption Solution is now joined by and the new Hardware Consumption Solution. A compute-as-a-service pricing model now incorporating both IBM Z software and hardware for the hybrid cloud.
In the Software Consumption Solution review, I note that it could motivate customers to move more apps onto IBM Z that require the highest security, availability, and transaction throughput. Customers transitioning to the new pricing offerings see benefits in greater flexibility in dealing with usage spikes and, most importantly, possible inclusion in a hybrid cloud architecture.
Most importantly, TFP Software Consumption does away with the practice of capping and makes all of the IBM Z infrastructure available.
Introducing the Hardware Consumption Solution for IBM Z
Customers wanted a better way to handle the more unpredictable shorter workload spikes throughout the day. The Hardware Consumption Solution provides a layer of always-on fixed capacity above the customers' base purchased capacity. If a workload spikes into the additional capacity (IBM calls it a "corridor"), usage charges will result. Costs include a subscription to rent and access the extra capacity and usage charges when consumed.
The amount of additional capacity that is available in that corridor determines the subscription charge. Essentially a percentage of the purchase price for that specific client. The more capacity, the higher the subscription cost, which can be prepaid or paid monthly.
The usage charge is measured in 15-minute intervals and charged at the peak MSU usage within that hour. Usage of the additional capacity for more than four hours in a day results in a daily charge. MSUs used measured by the SCRT (mentioned earlier) determine the usage charge.
Consumption-based pricing is available only for z15 generation and general-purpose central processors (CPs), with no specialty engines. Also, there must be a TFP Software Consumption Solution in place or planned for the software.
IBM will conduct a workload analysis to see if a customer qualifies for the offering. A workload analysis tool takes three months of data, looks at spikes in demand, and recommends the amount of additional capacity over and above the z15 base capacity the customer has already purchased.
With this pricing model, customers can leverage an existing investment into something more flexible—the available always-on capacity results in better efficiency, reduced overhead, and shorter response times. The less-than-a-day granularity will help alleviate the impact of sharp, unpredictable spikes common in modern workloads.
By having a new billing approach based on the amount of work done, rather than peaks reached, customers can expect jobs to finish faster yet at the same cost, and online processing can process more transactions simultaneously hence improving response times. Batch processing will also benefit, reducing batch windows dramatically.
On the journey to hybrid cloud, new types of workloads continue to emerge. High spiking business-critical workloads are becoming even more common, due in part to rapidly changing consumer behavior and unforeseen factors such as regulations and a global pandemic.
As a result, pricing flexibility and management are essential in a hybrid and multi-cloud world. The IBM Z now has a pricing model that delivers flexibility, simplicity, and more predictability across hardware and software. Compute-as-a-service pricing for the IBM Z again demonstrates IBM’s ability and willingness to listen to its customers and evolve the IBM Z to meet the changing requirements of the hybrid cloud.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.