Switching to CMP? Minimize Your Baseline Now
When moving batch workloads around to lower your R4HA, duplicate product peaks are a common challenge—and can cause their fair share of headaches. To remedy this issue, IBM recently announced a new pricing model—Country Multiplex Pricing (CMP). IBM has designed the new model to give even greater flexibility.
In short, it allows you to move and run workloads across your data centers in a single country with less financial impact than you’d experience by staying with your present VWLC model.
Country Multiplex Pricing (CMP) by the Numbers
Applying to z196, z114, zEC12, zBC12 and z13 processors, CMP measures a Multiplex, which is a collection of all these processors measured as if they were one machine. The SCRT creates one price per product based on MLC capacity growth across the Multiplex. This is intended to simplify the pricing process as well as hardware and software migrations.
Not only does CMP correct the issue of duplicate product peaks, but it also gives you the chance to treat your entire enterprise as if it is just one server. In essence, you can move work where and when you want while lessening the risk of a pricing increase.
Maximizing Your Savings
But there’s a gotcha. The initial MSU base will be computed based on a three-month average of the Multiplex peak for each product, not the SCRT peak. So, if you want to move to CMP, your goal is to first reduce that average as much as possible. If you’ve been playing around with trying to reduce your software costs, you already know this is difficult. You get it right one month, and the next month something changes and the price goes up.
You can avoid this problem. Before moving to CMP, preemptively use Automated Capacity Management (ACM), a feature of the powerful Compuware ThruPut Manager batch automation software. ACM reduces MSU consumption by constraining or deferring selected batch workloads. This means you can make a dent in the R4HA by defining the batch workloads you wish to have managed. Most people select low-priority batch as a starting point (usually about 25% of the workload contributing to R4HA peaks). This will ensure the best outcomes when you move to CMP.
The example below shows how a ThruPut Manager client was able to reduce consumption by about 17 percent using ACM. This reflects an immediate and ongoing savings of $163,100 per month, or almost $2 million annually.
In addition to making the transition to Country Multiplex Pricing easier, switching to ThruPut Manager means it’s there for you, even after the conversion. Let’s face it, workload management isn’t getting easier, despite simplification of pricing calculations. Even IBM acknowledges that. To optimize pricing without sacrificing performance, you need a higher level of workload management, and you can’t get reach a level higher than the automation of ThruPut Manager.
Latest posts by Denise Kalm (see all)
- How to Control Your CPU Demand and Costs at the Job Level - November 16, 2017
- Why ThruPut Manager Is the Real-time Cost-cutting Robot for You - November 9, 2017
- Batch Jobs Need More CLASS? Group Workloads by Service Groups - October 24, 2017