What are the requirements of taking advantage of the zIIP specialty engine?

What are the requirements of taking advantage of the zIIP specialty engine?

August 05, 2009 0 Comments

In this podcast, Gregg Willhoit explains the requirements for taking advantage of the zIIP specialty engine. The podcast runs for 2:53.

Click here for listen to the podcast:

Gregg Willhoit:

The requirements for a zIIP are pretty basic. You have to have a z9 or a z10. I think the base operating system is around 1.7, personally I would recommend 1.9. IBM has made it extremely cost effective to acquire a zIIP. There are numerous zIIP acquisition programs. IBM has provided the ability for the customer to, without even having a zIIP, determine if the zIIP can be cost effective. They do this through projected CPU parameters that can be set dynamically. It allows you to determine what percentage of certain workloads run zIIP eligible, and that can be key to justifying a zIIP.

IBM has various incentives for acquiring a zIIP, especially if you’re executing or using it with a new workload, which Shadow would be. IBM also, because the zIIP and the zAAP are similar, they’re basically the same engine, but they are limited in the types of workloads they support. They are somewhat fungible themselves. So if you’ve got a zAAP and no zIIP, and you need a zIIP, I’m pretty sure IBM can come out and switch your zAAP to a zIIP relatively easily.

So today there are almost no barriers to reducing mainframe TCO. And I think it’s incumbent on everyone to take a look at specialty engines, because the benefits that are derived are just incredible. It’s the first time in my career, and I’ve been doing this a long time, that I’ve seen such an offering – a hardware company which delivers such benefits to its user community. It’s just never happened before. And that’s why I decided to take full advantage of this as soon as possible. I wanted to give something back as well. It’s only going to bring excellent things to come in the future.

More and more vendors are going to offload more and more workloads. zIIPs and zAAPs will be more and more used. The mainframe will continue to not only just survive, but thrive, because of its inherent architectural advantages over non-mainframe platforms. And now the lower TCO afforded by the specialty engines from IBM. I mean it’s clearly the future’s so bright I’ve gotta wear shades situation for me, as far as I’m concerned.

Utilities can offload to the zIIP, but really the applications that can best take advantage of the zIIP are those which do consume significant amounts of resources. Utilities and monitors and things like that; it’s great that stuff can be offloaded to the zIIP. But if you’ve written a performance monitor that runs on z/OS and you’re getting significant zIIP offload in terms of CPU, you didn’t write a very good monitor in the first place, because the things are just not supposed to consume a lot of resources.

So the true benefactors of the zIIP are those workloads and applications which are driven by users that end terminals or B2B type operations. We’re talking extremely high transactional data volume that executes during the day. We’re talking about the stuff that truly drives the business of the organization. Those are the types of applications and workloads that most benefit.

Gregg Willhoit

View all posts from Gregg Willhoit on the Progress blog. Connect with us about all things application development and deployment, data integration and digital business.

Comments are disabled in preview mode.