How well is your mainframe outsourcer managing capacity and performance? – Part 1

There are many reasons why companies choose to outsource their mainframe. One of the major reasons is economies of scale—cost per unit (e.g. per MIPS or MSU) for hardware and software falls dramatically as the mainframe gets larger. Similarly, the cost of having the critical mass of specialist skills required to manage a mainframe is high, so it makes sense to be part of an environment where these resources are shared.

It is important to remember that outsourcing the work is not the same as outsourcing the responsibility. Companies choosing to outsource their mainframe still need to be able to follow up on their outsourcer to ensure they are getting what they pay for. The problem is that companies often end up lacking access to the required data and no longer have the skills to understand that data. They have lost the ability to follow up and the consequences can be expensive.

This is the first of a series of blogs where we will go through the basic knowledge and data that an outsourced customer needs to have in order to stay in control. This is not just about catching the outsourcer when they make mistakes, but also about building up a better relationship with the outsourcer based on mutual understanding and transparency.

This first blog focuses on why this is important. Why should you invest time and resources understanding how well your outsourcer is managing capacity and performance?

Coming blogs will focus will dig into more depth on topics like:

• The basic elements of CPU-based pricing for outsourced mainframes: Understanding MIPS and MSU
• The pitfalls of MIPS
• The mainframe outsourcer’s cost structure
• Billing models for outsourced mainframes
• The trade-offs between cost and performance.

Why is following up on your outsourcer important? Let’s take some examples:

Outsourcers make mistakes when invoicing for capacity.

For example, the technical people creating the basis for invoicing may not have read or understood the contract, and you end up getting billed for something other than what was agreed. You need to be able to crosscheck the data and the calculations. We have seen examples where outsourced customers have been overcharged hundreds of thousands of Euros a year for several years due to a disconnect between what was in the contract and what they were actually being billed for.

Outsourcers makes tradeoffs that affect your cost and quality of service.

For example, one way an outsourcer can ensure they live up to a Service Level Agreement about response time is to ensure that your LPARS have no resource constraints during peak hours. With no constraints, you may be paying to run lots of low priority work during your peak hour that could just as well run at some other cheaper time. But the outsourcer is happy. They are meeting the SLA and increasing their revenue at the same time.

If you don’t have a clear SLA, then the opposite might happen. If your peak happens to be in the same time as the outsourcers overall peak, they may limit your capacity in order to reduce their overall costs. This can have a serious impact on your quality of service at a critical time.

An outsourced environment is a shared environment.

Your costs and quality of service are deeply affected by the overall machine configuration and the behavior of the other customers on the machine. The outsource may make tradeoffs to reduce their own costs or help other customers at your expense. This is a fact of life when running in a shared environment, but it is important to understand what is going on and be able to have a fact-based dialog about it with your outsourcer.

You need to understand your cost drivers and what you can do about them.

Are you running low priority batch programs in your peak hours? Has an application change on your side resulted in a significant increase in resource usage? Which programs should you look at optimizing in order to save money? Is the outsourcer running system utilities like HSM during your peak hours? Do you have enough zIIP capacity, or are you running some zIIP eligible workload on much more expensive CP’s? Your outsourcer may not be very proactive in helping you answer these questions, especially if the answers will reduce their revenue.

Transparency is key to a good working relationship with your outsourcer.

A common understanding of what you are paying for, how it is measured, and what you and the outsourcer can do about it is critical for a good working relationship. Naturally the outsourcer wants to maximize their revenue and minimize their costs, but the good ones also think about maximizing the value they can create for their customers. Building a relationship based on alignment of interests requires transparency into the shared cost drivers.

Considering alternatives means understanding what you have today.

Understanding what you are paying for and how it is measured or calculated is a precondition for being able to compare alternative pricing models. This is true whether you are renegotiating your current agreement, considering moving to another outsourcer or just benchmarking your costs against the market. Are they both calculating MIPS in the same way? What software and services are included in the MIPS price? Which pricing model is most advantageous based on your current (and planned) usage patterns? What are the likely synergies that a new outsourcer will achieve taking over your workload?

Answering these questions requires access to data and the skills to understand that data. For one example of how SMT Data has helped an outsourced customer with these challenges have a look at this case study.

In the next blog we will dive into understanding what you pay for. This will include a basic introduction to CPU seconds, MIPS and MSU and some of the pitfalls that surround them.

Originally posted at SMT Data.