If you are an IT professional who works on IBM z Series mainframes, then you’ve probably heard about zIIPs and other “specialty processors.” But you may not really know what they are, what they do, and why they exist. With that in mind, let’s take a brief journey into the world of specialty processors.
Starting in the early 2000s, IBM has introduced several different types of specialty processors. The basic idea of a specialty processor is that it augments the main general-purpose CPUs. Instead of running all workloads on the general-purpose CPUs, specific workloads are shuttled to the specialty processors for execution.
Why is this useful or interesting to mainframe customers? Well, the specialty processor workload is not subject to IBM or (for the most part) independent software vendor (ISV) licensing charges. And, as anybody who has ever looked into mainframe software pricing knows, software cost can be many multiples more expensive than the hardware cost. Now mainframe pricing and licensing is complex and can be quite confusing, but at a high level, your organization’s monthly mainframe software bill is based on the peak average usage during the month.
Most mainframe software contracts are tied to the processor size of the machine on which the software is to be run, and the cost of the software rises as the capacity on the mainframe rises. But if capacity can be redirected to a specialty processor, then that workload is not factored into the software license charges. If enough workload can be redirected to specialty processors, meaningful cost savings can be realized.
Another benefit of the specialty processors is that they are significantly cheaper to acquire than general purpose processors. A standard mainframe CP can cost more than half a million dollars, whereas the list price of a specialty processor is about a quarter of the cost… and the street price of a specialty processor can be much less.
Specialty processors can be purchased for a one-time charge per engine, including no-charge replacement by faster zIIP engines when upgrading to a new machine. So many organizations today are augmenting their mainframes with specialty processors to delay costly upgrades.
But, of course, there is a catch! The specialty processors can only run certain types of workloads. There currently are three different types of mainframe specialty processors:
- ICF: Internal Coupling Facility – used for processing coupling facility cycles in a data sharing environment.
- IFL: Integrated Facility for Linux – used for processing Linux on System Z workload on an IBM mainframe.
- zIIP: Integrated Information Processor – used for processing certain, specific types of distributed workloads.
There used to be a fourth type of mainframe specialty processor, the zAAP, or Application Assist Processor. Its usage was designed specifically for Java workloads and XML parsing. However, late in 2009, IBM provided the ability for zAAP workloads to run on the zIIP, thereby enabling organizations to run zIIP- and zAAP-eligible workloads on a single type of specialty processor, the zIIP.
The ICF and IFL are designed for specific types of workload, dedicating coupling facility workload in the case of the ICF and processing Linux workload in the case of the IFL. By running these types of workloads on a specialty processor, the work will not apply to your monthly IBM software charges. The cost benefit is quite straightforward for these specialty processors. Although the zIIP offers a similar benefit, there are many nuances that need to be understood and considered.
Let’s Talk About the zIIP
The zIIP is a dedicated processor designed to operate asynchronously with mainframe general processors (GPs). When you activate zIIP processors, some percentage of the relevant workload can be redirected off of the general processors onto the zIIP specialty processor. The primary benefit of redirecting work to the zIIP is that IBM will not impose software charges on workloads that run on the zIIP.
Careful readers will note the phrase “relevant workload” in the previous paragraph. Not everything can run on the zIIP, only workloads that IBM deems as “new” are permitted. Originally, the zIIP was designed to support redirecting newer Db2 functionality, but over time the list of what is deemed to be “new” by IBM has grown. At a high level, the current zIIP-supported workloads include Java application programs, IBM z/OS Container Extensions (zCX), IBM Watson Machine Learning for z/OS, IBM z15 System Recovery Boost, and for some types of Db2 processing (e.g., XML, distributed queries, and some utilities). Other ISVs also have zIIP-enabled their products which enables portions of those workloads to run on zIIP processors. We will discuss the relevant workloads in more detail in an upcoming blog post.
There are limits to your usage of zIIPs that must be understood. First, there are limits on the number of zIIPs that can be installed. Originally, there could be no more than one zIIP per GP in a central processor complex (CPC). Today, some models allow two zIIPs per GP. Second, IBM’s license agreement places restrictions on the kind of code that is eligible to run on a zIIP; the code must run in a z/OS enclave under the control of an SRB (service request block). We will delve more into this topic in a later post, too.
Additionally, not all zIIP-eligible workloads will actually run on the zIIP. It can be troublesome to understand exactly what is being redirected, exactly when, and exactly how much of the workload is being redirected. Nevertheless, the primary intent of the zIIP is to reduce your IBM software charge, and the more workload that can redirected to the zIIP, the more your monthly cost savings can be.
Specialty processors are here to stay, and they can be used to help reduce your monthly IBM software license charges and thereby reduce the cost of mainframe computing. Although specialty processors introduce some complexity into management and capacity planning, organizations can benefit from exploiting them. And IBM, as well as ISVs, continue to introduce new offerings and functionality that can run on zIIPs, enabling organizations to better utilize specialty processors for more and varied types of workloads.
The bottom line is that an investment in specialty processors is likely to pay for itself in no time at all.
Craig Mullins, Copyright Craig Mullins Consulting, Inc