Critical Success Factors for Mainframe Capacity Management at Topdanmark
Topdanmark is a leading insurance company in Denmark offering a wide range of insurance products to private individuals, businesses and agriculture. The company employees over 2400 people and has been in the market since 1898.
Topdanmark’s IT system landscape
Topdanmark’s core applications are developed in a mixture of Cobol, Java, PL/I, Python, SAS and DB2 and are hosted on an outsourced mainframe. It is critical for Topdanmark’s business that these systems balance high reliability and performance with the lowest possible cost. Topdanmark uses SMT Data’s ITBITM for Z solution to help achieve that balance through effective capacity management and performance optimization.
In-house and Outsourced Operations
Topdanmark understands the importance of collaboration between in-house teams and external partners. While the company handles critical tasks such as software development, database administration, storage management, job scheduling, and release management in-house, they have outsourced the mainframe hardware and the IBM software licenses to Kyndryl. This strategic combination allows Topdanmark to leverage the expertise of external partners while maintaining control over operations that are crucial to their core business.
Topdanmark’s mainframe runs in an outsourced shared service environment. They have a flexible capacity management process with the outsourcer, Kyndryl. Capacity for the coming month is agreed in advance in terms of:
- Capacity Group Limit. This is a soft capping that allows Topdanmark to go over the limit for short periods but restricts their 4-hour rolling average (4HRA) MSU consumption to an agreed maximum.
- Absolute Capping. This is a hard limit that is set somewhat higher than the soft limit.
These limits can be adjusted on short notice, for an additional cost, if Topdanmark requires additional capacity beyond what was agreed during the monthly planning. This may happen for business reasons like a sudden storm increasing the number of insurance claims. It may also happen for technical reasons like a ‘rouge’ workload suddenly using more capacity than it should.
In September for example, Topdanmark had problems with the nightly batch, causing it to spill over into the peak online time. Topdanmark bought 200 extra MIPS so these batch jobs would have enough capacity to complete without affecting performance of business-critical online workload. Having complete performance and capacity data and ITBI to analyze the data are key to managing this kind of event effectively.
Figure 1: CP MIPS usage and limits for working hours on working days, showing the spillover of nightly batch on September 28 and the temporary increase in capping implemented to deal with the situation.
Critical Success Factors
For Topdanmark, successful capacity planning and management involves:
- Making certain that there is enough capacity to guarantee good response time for the business-critical online applications, especially during working hours.
- Ensuring that the nightly batch has enough capacity to complete at least 4 hours before the morning online peak (due to the rolling 4-hour average capping).
- Keeping costs as low as possible by not ordering more capacity than is needed during the coming month.
- Avoiding expensive ‘ad hoc’ requests for additional capacity on short notice, unless driven by business needs.
- Using capping and WLM policies to ensure that low priority workloads get delayed if resources are constrained, allowing critical applications to get the resources they need.
- Quickly identifying any performance or capacity anomalies, understanding the root causes, and fixing them.
- Continually focusing on identifying and optimizing the top resource consumers during peak hours in order to reduce costs over time.
- Understanding the capacity and performance impact of new applications being rolled out, or existing applications being changed or re-architected.
- Correctly budgeting long term capacity needs and costs based on historical trends, seasonalities, and expected business growth.
Making it happen
Meeting these critical success factors requires access to detailed data about capacity and performance including lots of historical data. Good tools to analyze and visualize the data, and expert assistance to understand the results also play an essential role. This is where SMT Data comes into the picture with their ITBI for Z solution and their skilled consultants.
With ITBI and SMT Data’s support, Topdanmark has the critical success factors for capacity management under control and ensures stable operations with good performance. At the same time Topdanmark has continually reduced the yearly capacity costs, even during periods with growing business volumes.
Elise BundgaardSolution Manager at Topdanmark