How Can I Establish a Solid Foundation for Intelligent Mainframe Capacity Management?

SOLUTION BRIEFDynamic Capacity Management - CA Technologies

DRAFTSOLUTION BRIEFCA DATABASE MANAGEMENT FOR DB2 FOR z/OS

CA Dynamic Capacity Intelligence from

CA Technologies is an easy-to-use, fully automated

and intelligent capacity management and real-

time dynamic capping solution. The solution is

designed to streamline your ability to monitor and

effectively manage SLAs and improve operational

efficiency. It helps IT to achieve rapid value by

enabling priority mainframe workloads to the

required capacity to meet SLAs while at the same

time helps remove the guesswork for MLC pricing.

ChallengeAny breach of Service Level Agreements (SLAs) for priority workloads is unacceptable. The dilemma: IT must provide the best possible service to the business, while the business must prioritize cost containment strategies. It's a paradoxical situation.

OpportunityIT and business stakeholders desire a predictable and proactive way to help provide that critical workloads complete per SLAs with minimal impact to cost. Selecting the best possible combination of pricing options, monitoring variables and resource management strategies can lead to significant savings.

BenefitsCA Dynamic Capacity Intelligence, for z Systems®, provides proactive, predictable capacity management for optimizing system resources for prioritized workloads. This capacity management tool helps you better manage and utilize mainframe capacity, thus helping drive down Monthly License Charges (MLC). Intelligent automation avoids unplanned spikes in cost and empowers IT to better manage Service Level Agreements (SLAs) through automated, dynamic capacity optimization. It helps you:

Improve management of SLAs. Enable continuous optimization of missioncritical application delivery and operations.

Automate continuous optimization for priority workloads. Shift availablemainframe capacity across LPAR boundaries.

Contain and control software costs. Flexibility to help provide that criticalworkloads complete per SLAs with minimal impact to MLC.

Executive Summary

3 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Take the Guesswork out of MLC PricingIn many environments, breaching SLAs of priority workloads is unacceptable. IT knows they must provide the best possible service to the business. Conversely, businesses prioritize cost containment strategies because exceeding the budget is viewed as almost equally unacceptable as a breach of SLAs.

Debating between violating capacity limits with increasing cost, or impacting business-critical SLAs, becomes heated and sometimes leads to finger pointing while your customer experience suffers. The challenge becomes how to meet expected SLAs at the best possible price.

MLC can be easily considered the largest single monthly invoice for businesses where mainframes are deployed. MLC charges are determined by 4-hour peak-capacity-usage during a month (called “Rolling 4-Hour Average” – R4HA.)

Both IT and business stakeholders desire a predictable and proactive way to help provide that critical workloads complete per SLAs with minimal impact to cost. Selecting the best possible combination of pricing options, monitoring variables and resource management strategies can lead to significant savings.

Intelligent Automation Improves Management of SLAs and Avoids Unplanned Spikes in Cost.CA Dynamic Capacity Intelligence is an easy-to-use, fully automated and intelligent capacity management and real-time dynamic capping solution. The solution is designed to streamline your ability to monitor and effectively manage SLAs and improve operational efficiency. It helps IT to achieve rapid value by enabling priority mainframe workloads to the required capacity to meet SLAs while at the same time helps remove the guesswork for MLC pricing.

It recognizes the difference between urgent, time-and business critical tasks and lower-priority tasks that can be briefly postponed. The solution is designed to help lower the overall Million Service Units per hour (MSU) usage during peak cost times by automating mainframe capacity balancing to reduce MLC charges while helping to achieve all time-critical workloads.

It provides a single point of control [Figure 1] to simplify and automate enterprise-wide deployment, monitoring, and reporting to insulate staff from the underlying complexities of many aspects of their operations, including the overall spectrum of software portfolio usage and cost.

4 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Why CA Dynamic Capacity Intelligence?

Unlike other optimization methods available today that move workloads to larger capacity, the solution helps customers maintain SLAs for critical workloads by continually analyzing workload capacity usage and automatically and dynamically move capacity to where and when it is needed.

The solution is very flexible, having the ability to work either with, without, or within IBM Group Capacity Limit (GCL, for groups of LPARS) and Defined Capacity (DC, for singular LPARs).

You can prioritize MSU capacity based on workload priorities within an LPAR or across LPARs and lets you define a flexible range for MSU usage and automatically allocates the minimum usage.

If defined minimum usages have already been exceeded, they can be defined to become the new baseline.

By shifting more resources to deliver business innovation and lowering software cost across mainframe operations, you’ll achieve better business continuity expected by your CTO and better deliver the mainframe platform economics demanded by your CFO and improve your overall customer experience.

Figure1: Easily view your peak real-time rolling four-hour (R4HA), Caps and MLC.

5 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

6 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Establishing a Successful Foundation for Intelligent Mainframe Capacity ManagementThe ultimate goal is to empower IT to better manage service level agreements through automated, intelligent mainframe capacity management thus helping drive down MLC. But what are the top challenges you may face when implementing SLA and cost controls? How can CA Dynamic Capacity Intelligence help you conquer those challenges? Let’s investigate those challenges and how CD Dynamic Capacity Intelligence can help you conquer those challenges.

Understanding Your Monthly License ChargeIBM® charges enterprises for its proprietary software such as z/OS®, DB2®, CICS,® IMS®, WebSphere MQ®, etc. based on the highest usage over any given four-hour period (i.e. R4HA) during the monthly billing cycle. These charges can usually be significantly lowered by allocating the available resources more efficiently and considering a number of other impacting factors.

Where might your enterprise be overpaying for unneeded capacity? What can you do to address each of these impacting factors? The following pages will go through each of the factors listed below to guide you and your enterprise through the potential challenges ahead.

Billing periods do not match peak usage.

Default control mechanisms GCL and DC are inadequate and inflexible.

No capacity balancing across LPARs based upon Workload Priorities.

R4HA is slower than Actual MSU Usage.

Capping is "bad".

Capping values are increased in large steps.

Knowledge drain-technical and pricing.

Increasing IT Complexity.

No real-time information: Does capping really offer a financial reward?

7 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Billing Periods Do Not Match the Peak Usage

MLC billing periods start at the second day of a month and include the first day of the next month. The disadvantage for many enterprises is that their highest system usage occurs on the first working day of the next month, due to end of month processing for the previous month. This day is also the last day of the billing period.

Whenever enterprises have tried to save MSU during the month and are disappointed at the last day of the billing period, chances are that after experiencing this several times, they give up and set caps too high, and in the end not saving any MSUs.

Using software like CA Dynamic Capacity Intelligence to define a flexible range for MSU usage and to automatically allocate the minimum usages, making sure that the MSU usage over time is as low as possible and as much as needed, but not exceeding the defined maximum minimizes the impact special MSU hungry events such as month end processing.

Default Control Mechanism GCL and DC are Inadequate and InflexibleTo control the usage, IBM offers Defined Capacity (DC) for single LPARs and Group Capacity Limit (GCL) for groups of LPARs. These mechanisms offer no flexibility as they need to be set and reset manually. Basically it is only helps protect aganist very high usage. Whenever a GCL is reached, all MSU usage is divided based upon previously defined settings. They very rarely reflect the real need during the actual peak time.

The only alternative is to define a flexible minimum and maximum MSU range that automatically allocates the minimum usages required to meet the workload.

Capacity should only be added when it is necessary in order to run the critical workload. This approach makes sure that the MSU usage over time is always as close to the required minimum as possible, while not exceeding the defined maximum. CA Dynamic Capacity Intelligence continuously monitors MSU usage and provides additional capacity where needed without exceeding defined maximums.

No Capacity Balancing Across LPARs Based Upon Workload PrioritiesCapping is done per single LPAR (Defined Capacity) or per Group of LPARs (Group Capacity Limit) based upon the R4HA maximum. However, the goal is to use the available capacity optimally so it is necessary to always monitor if one LPAR is struggling to execute the Time-Critical workload while at the same time another LPAR still runs workload that is not time critical.

Different LPARs may run similar workloads in identical workload classes but with a different level of time criticality such as batch workload in production vs. batch workload in development. The level of time criticality may even differ depending on the time of day such as production batch workload during the day and during the night).

Use software such as CA Dynamic Capacity Intelligence in conjunction with z/OS Workload Manager (WLM) to define the levels of time criticality per WLM Importance Level. Further, define multiple capping policies, each activated, depending on the day/date and time of day, not time critical workload is slowed down to avoid having to add more capacity.

8 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

R4HA is Slower than Actual UsageAt many sites, the actual MSU peaks over a two-hour period only with a rapidly declining usage from just before lunch time onward. Because the R4HA needs to reflect the usage of the past four hours, it is much slower than the actual MSU usage. In peak situations of less than four hours flat system usage, the result is that the R4HA still goes up while the MSU demand has dropped. The result is that enterprises are paying for capacity they do not use.

Automatically perform "capping without limitations" with CA Dynamic Capacity Intelligence. If the current usage, Actual MSU (red), drops below the R4HA (green), the Defined Capacity (blue) follows the current usage [Figure 2].

There are three advantages to this method:

Users will not notice the capping as there is no limitation. The Sub Capacity Reporting Tool (SCRT) will take the lower of the two values of R4HA and DefCap. Capacity that is freed up may be used by other LPARs at no additional cost.

Figure 2: Shows the relationship amongst actual MSU, R4HA and CA Dynamic Capacity Intelligence

Capping is "Bad"Defined Capacity (for single LPARs) and Group Capacity Limit (for groups of LPARs) limit the MSU usage. They are simple ways to make sure that the MLC costs do not exceed the defined maximum. They are not MSU management tools!

When used too aggressively, not monitored closely or used in an environment in which the system settings are not optimal, problems will occur during peak time. For example, online transactions may be slowed down while at the same time low-level batch workload still runs fine. Many enterprises refuse to cap their production workloads in any way.

Caps need to be recalculated regularly to guarantee that there is enough capacity to perform all time-critical workloads. Use CA Dynamic Capacity Intelligence to recalculate your caps every five minutes. In addition, the time-critical workloads need to be monitored permanently to guarantee they perform well.

9 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Capping Values are Increased in Large StepsIn many cases the DefCap or GCL is raised considerably when LPAR or group capping threatens to be activated. Time is of the essence to avoid capping and often sysprogs lack the information to do more than make an educated guess

With automated capping, the increases are calculated every five minutes. This means that increases only have to cover usage needs for the next five minutes. These small MSU increases guarantee that the time¬ critical workload gets done as scheduled.

In Figure 3 below on the left, the Defcap was manually raised from 490 to 550 at around 10:30 a.m. and set back to 490 at 4:30 p.m. The R4HA did not come close to the value of the defined capacity, leaving plenty of capacity for discretionary workload. On the right the automated capping, showing that the DefCap is raised according to the needs, just before capping kicks in. At the same time , the not time-critical workload is slowed down.

As soon as the actual usage drops below the R4HA line, an automated capping tool like CA Dynamic Capacity Intelligence will intelligently manage capacity - not add additional MSUs to the DefCap. As a result some low-level workload may be executed and the DefCap value will be reduced again later. In this example, it leads to savings of around 70 MSU all without impacting critical workload. Assuming this enterprise runs IBM's z/OS, DB2, CICS and NetView, the savings are likely to be more than US$400 per MSU, or in excess of US$28,000 for this one month.

Figure 3: Automated DefCap usage can help reduce MSU usage much more effectively than manually

Knowledge Drain - technical and pricing

z Systems programmers are quickly becoming an increasingly scarce resource that cannot be easily replaced. Each seasoned mainframe professional who retires or departs an enterprise leaves behind a major knowledge gap.

Use the z/OS WLM feature and CA Dynamic Capacity Intelligence and put the MSU settings into intuitive software parameters and document these so that the technical staff can simply monitor the performance of an automated program. This provides two benefits:

Frees up experienced staff to attend to potentially more vital tasks. Facilitates training new staff.

10 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES ca.com

Increasing IT ComplexityAs new platforms are interlinking with each other and using z Systems as a major data hub and repository for massive data volumes, many systems become too complex and difficult to manage manually. Allocating MSU capacity optimally over small environments is a challenge, but doing it optimally for large environments is simply an impossible manual task.

Define parameters and rules for each LPAR, depending on the LPAR specific workload (Production, Development, Sysprog, etc.) and specifications (Online only, Batch LPAR, etc.) and let software such as CA Dynamic Capacity Intelligence automatically calculate the optimal settings every five minutes.

No Real-Time Information: Does Capping Really Offer a Financial Reward?In many environments development LPARs are limited in their MSU usage. Numerous situations have been found where there was no, or almost no, financial gain to capping. For example, during times that production LPARs do not need much, some of that excess capacity could be given to development LPARs, giving application development more capacity at no or only very little extra costs.

CA Dynamic Capacity Intelligence can automatically solve the challenge of knowing real-time if limiting the MSU usage has a financial impact through its LPAR-specific parameter settings.

ConclusionCA Dynamic Capacity Intelligence provides proactive, predictable capacity management for optimizing system resources for prioritized workloads. It:

Helps provide that priority workloads complete per SLAs with minimal impact to MLC. Moves available capacity dynamically to LPARS running critical workloads that need it. Automates in real-time the capacity balancing to fully utilize available capacity. Helps decrease the manpower that is linked to real time capacity management. Provides complete transparency into actions taken when moving capacity.

CS200_SPK_0622

11 | SOLUTION BRIEF: DYNAMIC CAPACITY MANAGEMENT - CA TECHNOLOGIES

CA Technologies (NASDAQ: CA) creates software that fuels transformation for companies and enables them to seize the opportunities of the application economy. Software is at the heart of every business, in every industry. From planning to development to management and security, CA is working with companies worldwide to change the way we live, transact and communicate – across mobile, private and public cloud, distributed and mainframe environments. Learn more at ca.com.

Copyright © 2017 CA. All rights reserved. All trademarks, trade names, service marks and logos referenced herein belong to their respective companies. This document is for your informational purposes only. CA assumes no responsibility for the accuracy or completeness of the information. To the extent permitted by applicable law, CA provides this document “as is” without warranty of any kind, including, without limitation, any implied warranties of merchantability, fitness for a particular purpose, or noninfringement. In no event will CA be liable for any loss or damage, direct or indirect, from the use of this document, including, without limitation, lost profits, business interruption, goodwill or lost data, even if CA is expressly advised in advance of the possibility of such damages. CA does not provide legal advice. Neither this document nor any CA software product referenced herein shall serve as a substitute for your compliance with any laws (including but not limited to any act, statute, regulation, rule, directive, policy, standard, guideline, measure, requirement, administrative order, executive order, etc. (collectively, “Laws”)) referenced in this document. You should consult with competent legal counsel regarding any Laws referenced herein. Some information in this publication is based upon CA’s experiences with the referenced software product in a variety of development and customer environments. Past performance of the software product in such development and customer environments is not indicative of the future performance of such software product in identical, similar or different environments. CA does not warrant that the software product will operate as specifically set forth in this publication. CA will support the referenced product only in accordance with (i) the documentation and specifications provided with the referenced product, and (ii) CA’s then-current maintenance and support policy for the referenced product.

Connect with CA Technologies at ca.com

To learn more about our mainframe capacity management technology, and how it can help you achieve real savings in your environment, please visit us today: ca.com/dci