Consumers appreciate the convenience mobility offers. They can do business from anywhere, at any time, using whatever devices they desire. Mobility, however, offers challenges for mainframes. While mainframes have successfully provided consistent performance and availability—achieved with rules and thresholds to quickly identify when something starts to go wrong—mobility introduces new, less-consistent activity patterns to mainframe applications. This includes a dramatic increase in transaction volumes, users, and devices accessing applications nonstop. With such unpredictability, the old rules for how to manage and respond must change.
The dynamics of these mobile application behaviors create a transition from the “typical” monitoring scenario of one threshold over a “shift” of time to a more dynamic monitoring behavior. This move requires solutions to learn behavior patterns and adjust and maintain multiple thresholds over the same shift of time to quickly detect anomalies.
It’s critical to business success to learn the new activity patterns driven by mobility and to use that information to modify the rules for managing mainframe performance.
Understanding the Challenges
Mainframe workloads used to be predictable. In banking, customers handled their transactions during the day at a local branch. The IT department ran batch jobs at night to process deposits and checks, and to update customer accounts. The IT staff knew that during a specific period of time a particular application might process a few thousand transactions a minute.
Today, there aren’t many transactions a consumer can’t conduct from a mobile device. Consequently, mainframes are running considerably more applications accessed by many more people doing a greater range of transactions and using multiple devices. The applications are more complex as well.
IT also must meet Service Level agreements (SLAs) for normal workloads and for special situations. Let’s say an insurance company has a special offer on boat insurance. This promotion will cause an increase in the number of quote requests for that application subcomponent. IT must be prepared with sufficient capacity, because if response time is too slow, consumers may purchase elsewhere.
You can manage this challenge by breaking it down into these five steps:
1. Evaluate your applications. Effective monitoring starts with an in-depth understanding of each application and its subcomponents. Evaluate your applications by understanding their flow across your IT infrastructure and their roles in supporting the business. This insight will guide you in making sound decisions regarding the level of granularity you need for monitoring those applications.
2. Become more proactive. Implement an automated self-learning and adjusting solution to monitor the dynamics of the application workloads to learn of changing conditions. If response time for a subcomponent rises, your monitoring tool should alert the right people so they can step in before performance degradation occurs.
3. Automate ticketing. When you integrate incident management and monitoring systems, tickets are generated and routed automatically based on conditions your monitors detect. Automation eliminates manual ticket creation and speeds mean time to repair by accelerating the delivery of alerts to the right people.
4. Prioritize response. Swift response is important, but so is responding to issues based on business impact. For example, claims adjusters for an insurance company need access to an online application for submitting claims in the field. It’s an important application, but it isn’t as critical as the application that delivers insurance quotes and lets consumers purchase online. Performance degradations for consumer applications can translate into lost revenues.
5. Automate response. As the number of applications and the volume of users continue to rise along with the dynamics of each, IT can’t respond by adding staff. Instead, IT must take advantage of self-learning and adjusting technologies to intelligently automate actions so the current staff can handle the expanding workloads.
Taking the steps described here enables the current staff to proactively monitor and take full advantage of automation tools to meet the business needs without increasing the bottom line.