Soft-capping Versus Hard-capping: Part One
Are you taking advantage of IBM’s sub-capacity pricing models? One decision you’ve had to make is whether to accelerate your software savings with hard-capping or soft-capping. Both of these options come with pros and cons, and the one that works best for you will depend on your company’s business needs and your data center environment.
To help you decide, this two-part series will explore the details of each. This installment offers a rundown of hard-capping, while the second post will cover soft-capping considerations.
What it is: A hard cap is set by specifying the initial capping option in the LPAR definition. Basically, this means the LPAR is capped at its defined weight. It can’t exceed its guaranteed share, even if the CEC has capacity that is unused.
Who it’s for: It’s for service providers who want to limit the consumption of individual LPARs regardless of available cycles.
Other notes: Because a hard cap prevents an LPAR from exceeding a predetermined usage amount, even a quick spike in CPU requirements will result in delayed workloads. This greatly limits the responsiveness of the LPAR. Because the R4HA is an average, occasional spikes that exceed the desired CPU usage are okay, provided the rest of the time the LPAR is using less. This usage pattern is very typical—and not suited to a hard cap.
If your organization has implemented the hard-capping model, what advice would you give someone who’s thinking about doing the same?
Is soft-capping more your preference? In the second post of this two-part series we explore the pros and cons.