zDynaCap: Same Service at Lower MLC Costs thru Capacity Balancing
Many different factors impact how Workload License Charges (WLC) calculates the monthly IBM mainframe software invoice. There are two main approaches to lowering this cost. One is to use IBM’s Workload Manager to distribute workloads more efficiently, and the other is to use zDynaCap to distribute MSU capacity more effectively. These two products together provide a new magnitude of savings beyond what Workload Manager can do alone. zDynaCap makes sure that peak time CPU (MSU) capacity is optimally allocated while keeping high priority workloads operating at peak efficiency. The strategy is simple; allocate the MSU’s to where they are needed.
zDynaCap: The Technology
The core technology of zDynaCap features:
- Uses only IBM supplied z/OS functions and interfaces
- Has an expert system engine
- Proven scalability, reliability and efficiency
- Drill down reporting levels for executive, middle and line users
- 24×7 operation with system checks every five minutes
zDynaCap: The Background
The z/OS Mainframe user can somewhat control MSU usage by deploying soft-capping techniques, generally associated with Defined Capacity (DC) and Group Capacity Limit (GCL) functionality, and workloads can be prioritized via the Workload Manager (WLM). However, these controlling mechanisms lack the granularity needed by various mainframe sites to deploy policies yielding an optimal cost/performance balance. All too often, sites do not attain the lowest WLC cost, even when the workloads might appear constrained. If the MSU resource is not available for the mission critical high priority workload, then the associated z/OS WLC costs will likely rise. The only constant rule for Sub-Capacity Reporting Tool (SCRT) invoice generation is that the customer MSU metric is based upon the Defined Capacity (DC) or Rolling Four Hour Average (R4HA), which ever value is the lowest.
Mainframe workloads are typically comprised of a daytime on-line and night time batch window. The workload characteristics are vastly different for these two operations. When allocating the CPU resource, different MSU values are needed to accommodate the short-running TP (e.g. CICS, IMS) monitor transaction from the long-running batch job. Ideally there would be a CPU (MSU) resource allocation policy for each of these timeframes. Further, peak workloads are generated periodically (e.g., weekly, monthly, annually) and must be accommodated. The 21st Century “on demand” 24×7 transaction oriented workload peaks can also occur at any time.
So, an ideal solution to the problem must automate and dynamically apply a CPU (MSU) resource allocation policy. This process must be proactive and safeguard the high priority workloads without impacting the R4HA, whose peak values dictate the monthly SCRT invoice.
zDynaCap: Features and Functions
- A plug and play product with no hooks or interfaces other than the standard z/OS interfaces
- Automatically balances MSU capacity via customer defined policies (LPAR groups)
- Flexible enough to control either total costs or even product costs, the latter by defining LPAR groups based on product usage
- Control granularity is provided by nested group functionality
- Delivers a complete overview and control of MSU usage
- Function rich browser based reporting feature delivers a cockpit view of current usage information
- Reports and historical information offer cross-checking of the SCRT reports and associated invoices
- Client SCRT and WLM data (SMF Type 70 & 72) are used to:
- quickly determine a conservative savings forecast, including an RoI calculation
- setup a customized run environment based on user preferences
zDynaCap: Capacity Balancing
zDynaCap automates capacity balancing within CPCs, Capacity Groups or user defined Groups of LPARs. Central to zDynaCap are the user defined balancing policies. Within these balancing policies, users define the MSU ranges of the defined Groups and LPARs and also the priorities of the associated LPAR Workload. zDynaCap continually monitors overall usage and compares this to the available capacity and the user defined MSU balancing policies.
When required, zDynaCap re-allocates MSU capacity assigned to workload that is not time critical to workload within the same Group that is time critical. This optimizes MSU resource allocation and minimizes the associated R4HA, while at the same time guaranteeing that business critical applications get the capacity that is needed to run. Only if there is no ‘leftover’ capacity to be re-allocated within the defined Group and Time Critical Workload threatens to be slowed down, will zDynaCap add MSUs within the defined range.
Changes in MSU capacity values are implemented by accessing the HMC through the official IBM BCPii interface. HMC values set by zDynaCap can at any time be overwritten temporarily or permanently via the ISPF interface.
zDynaCap is easy to use and as it comes with an initial user specific set of balancing policies.
The functionality of zDynaCap is unique to IBM’s System z platform and is far superior to the limited functionality of DC or GCL. It normally leads to fast and easy savings, making zDynaCap pay for itself within a very short time.
zDynaCap balances MSU capacity within LPAR groups to guarantee that during peak time the mission critical workload is processed according business needs for the lowest possible MLC cost.
zDynaCap: Simplified MSU Usage & Policy Reporting For The z/OS Professional
All too often there is too much data and not enough information. zDynaCap provides an efficiently intuitive GUI interface that displays the information required by the z/OS technician who manages MSU capacity usage. This sample report shows the policy flexibility available for defining an MSU policy with calendar (e.g. Day, Time) & zServer (e.g. CPC, Group, LPAR) metrics:
zDynaCap: Combined Technical Analysis & Management Summary Reports
“A picture is worth a thousand words” is high on the list of desirable features for a present day product. Data Center Management requires compelling graphical reports to quickly assess how the operation is running. Knowing how the business policies are performing allows for quick interactions with the frontline technicians (E.g. Systems Programmers, Capacity Managers) and Senior Management (E.g. Director, CIO, CFO). zDynaCap performs the required data analysis to produce information rich drill down reports to satisfy multiple levels of need.
The GUI reports are browser based and can therefore be shown on Smartphone, Tablet and PC. They highlight the potential to refine the MSU usage policy, allowing mission and time critical workloads to process without being impacted by low priority workloads.
zDynaCap: Complementary Customer Workload Analysis
By analysing business specific SMF Type 70 (SCRT) and Type 72 (WLM) data, it is possible to very quickly highlight what MSU and associated z/OS WLC software savings can be achieved with zDynaCap. Together with the SCRT report, the pre-analysis allows for the delivery of a pre-customized software installation package. Sample policies are tailored to client’s environment. The only technical pre-requisites is that customers need to run z/OS 1.11 or later. zDynaCap is typically installed and running in several hours.
zDynaCap adds value and provides favorable short-term ROI and long-term TCO cost reduction:
- Bottom Line: The reduced IBM Software License Invoice gives the real dollar savings every month
- Technical Line: Optimized MSU usage
- Performance Line: WLM interaction safeguards that Mission & Time critical workloads process
- Reporting Line: Enhanced MSU usage policy definition reporting simplifies z/OS technical management & reporting
- Management Line: Reports allowing senior management to assess business performance
- Growth Containment: More efficient operations reduce zSeries resource acquisition (E.g. CPU upgrades)
- Staff Line: exponentially enhances the ability of experienced technicians to extract more throughput from their systems
The 21st Century business is compelled to do more with less and zDynaCap provides a mechanism for IBM Mainframe users to derive maximum return from their z/OS software financial investment.