Explaining Mainframe Pricing to the Cloud Guys

In as of late of hybrid cloud working with some purposes working finest in a cloud surroundings and a few that work finest on the mainframe, there’s going to come back a time when mainframers and cloud persons are going to be sitting in the identical room and the best way computing is paid for shall be mentioned.

The cloud guys will say pricing on the cloud is simple and also you solely pay for what you employ. They could additionally say that transferring to the cloud means there shall be a value discount as a result of much less bodily house is required, much less {hardware} is required, there’s a discount in pricey high-end software program, and fewer help personnel are wanted. They can even say that there’s all kinds of entry to the newest expertise, plus a standard framework throughout the applying infrastructure. As well as, ‘shelfware’ might be prevented since you use solely what you want. They might additionally recommend that there are benefits when it comes to safety. For instance, infrastructure should adjust to trade requirements, and information is much less liable to worker theft. By way of reliability, there’s built-in redundancy, and most suppliers assure 99.99% uptime. Plus, cloud technical expertise are plentiful making workers cheaper.

The mainframe individuals may recommend that with the cloud there might be price creep when it comes to computing assets and departmental use/abuse. There may be very prone to be downtime ensuing from communication failure and Web drops. There are safety points. It doesn’t guard in opposition to weak digital safety strategies. Maybe most importantly, there are efficiency points on the subject of conventional batch workloads and sophisticated transactions if they’re migrated to the cloud. There’s additionally a problem with cloud vendor lock as a result of distributors extremely encourage their very own merchandise, making it troublesome to maneuver off any specific platform, as soon as you might be there. 

However, the actual problem comes when mainframers attempt to clarify to the cloud guys how charging on the mainframe works. Let’s take a look on the sorts of issues they should clarify.

Firstly, there’s MIPS (Hundreds of thousands of Directions Per Second), which, traditionally, was used as a measure of the variety of directions that could possibly be processed in a second of computing time, and was used to measure common computing capability. 

Then there are MSUs (Million Service Models), that are a measurement of the quantity of processing work that may be carried out in an hour. One ‘service unit’ initially associated to an precise {hardware} efficiency measurement, however that’s now not the case. A service unit is an imprecise measurement, and 1 MSU is roughly 8.5 MIPS. IBM publishes MSU rankings for each mainframe mannequin. 

One other downside for cloud individuals is the distinction between licensing and pricing. When you execute IBM zSeries software program on a CPC (Central Processor Advanced), you should have a license to take action. A license for an IBM month-to-month license charged product is restricted to the product and a CPC with a specific serial quantity. The license is offered when it comes to MSUs. In case you are executing a product on a 1500 MSU CPC, you should have a 1500 MSU license, specifying the serial variety of that CPC. The worth for a product, ie how a lot you pay IBM every month, will depend on the pricing metric that’s used for that product.

As soon as they’ve grasped that, you possibly can transfer on to Month-to-month License Cost (MLC) merchandise, which embody z/OS, Db2, CICS, IMS, MQSeries, and COBOL. The pricing and phrases and circumstances for MLC merchandise are based mostly on the pricing metric chosen. Pricing metrics can roughly be grouped into two classes: full capability and sub-capacity.

Underneath a full capacity-based metric, all software program expenses are decided by the capability of the CPC through which the product runs. Parallel Sysplex License Fees (PSLC) and zSeries Entry License Fees (zELC) are examples of full capacity-based metrics.

Underneath a sub-capacity metric, software program expenses for sure merchandise are based mostly on the utilization capability of the LPARs through which the product runs. Workload License Fees (WLC) and Entry Workload License Fees (EWLC) are examples of sub-capacity-capable pricing metrics. 

IBM lately launched Tailor-made Match Pricing (TFT), which is designed to be extra like cloud pricing. It’s based mostly on the general {hardware} consumption, not LPAR rolling four-hour common (R4HA). A baseline of total MSU consumption for the yr is taken together with an settlement to devour extra MSUs within the upcoming years. A reduced charge for MSUs is utilized, and month-to-month billing is predictable and the identical no matter month-to-month variations in utilization.

The costs for merchandise that use sub-capacity pricing are based mostly on how a lot the LPARs through which the merchandise run make the most of system assets, relatively than on the total capability of the CPC. Customers should purchase {hardware} capability for future wants with out incurring a direct enhance of their software program invoice. If the utilization decreases when enterprise is gradual, the software program invoice decreases with it. If utilization is seasonal, the month-to-month software program payments are decrease during times of decrease utilization. Customers pay for capability on a rolling four-hour common, not on the utmost capability reached. 

The rolling four-hour common represents the common consumption (in MSU) of the LPAR over the last 4 hours. Each 5 minutes, IMSU (Instantaneous consumption of MSU for the LPAR) is measured. The R4HA is a mean of the previous 48 IMSU metrics. This isn’t product usage-based pricing; as a substitute, it’s a little bit of a hybrid.

Outlined Capability (DC) is about within the {Hardware} Administration Console (HMC) and is used to regulate billing. It doesn’t implement capping and isn’t obligatory. The z/OS Workload Supervisor (WLM) calculates and displays the R4HA. If DC is about to a non-zero quantity, WLM displays the R4HA and ensures that the R4HA is lower than or equal to the DC. In fact, WLM additionally manages your workloads based mostly on settings/objectives. The Processor Useful resource/Programs Supervisor (PR/SM) enforces the gentle cap when WLM determines it’s wanted. 

The Sub-Capability Reporting Instrument (SCRT), a no-charge IBM device, experiences required license capability for sub-capacity-eligible merchandise. The SCRT signifies the required license capability (in MSUs) of every sub-capacity-eligible product. The SCRT cross-references LPAR utilization and product execution by LPAR to find out the utmost concurrent LPAR four-hour rolling common utilization – the very best mixed utilization of LPARs the place every product executes in the course of the reporting interval. Sub-capacity merchandise are charged based mostly on the rolling 4-hour common utilization of the LPARs through which the sub-capacity merchandise execute. The sub-capacity report determines the required license capability by analyzing for every hour within the reporting interval: the four-hour rolling common utilization, by LPAR; and which eligible merchandise have been energetic in every LPAR. 

Every hour, the R4HA is in comparison with the outlined capability (DC). If the DC is about, the SCRT makes use of the decrease of the 2 values for the utilization worth for the z/OS system for that hour. 

The SCRT makes use of SMF 70-1 and SMF 89-1 / 89-2 information. The SCRT is computed on a month-to-month foundation, from the second day of the month at 0h00 to the primary day at 24h00 of the next month. R4HA averages are calculated for every MLC product, every hour, for every LPAR, and for the month. 

The gentle capping rule is that when the R4HA turns into larger than or equal to the DC, the LPAR is capped. That implies that the IMSU consumption won’t be able to exceed the DC anymore till the R4HA turns into decrease than the DC. 

Let’s see how a lot of that the cloud guys perceive the primary time you inform them!