To say smart phones, which now account for 50 percent of all mobile phone purchases, have radically changed the way consumers interact with the world, would be an understatement. The average mobile phone user accesses his/ her device 30 times per hour to check news, post updates and photos, text, email, tweet, talk (yes, talk), watch videos, bank and shop online all while keeping abreast of what’s happening at work and home.
Just as our devices are available to us 24/7, 365 days of year, we require similar availability of the systems we access, and that includes the mainframe. If those systems aren’t available at the time and pace we want them, the consequences for business can be dire. Take, for example, a millennial bank customer who checks his balance from his mobile phone at 2:30 a.m. only to find his balance is significantly lower than what he thinks it should be. Naturally, he vents his concerns and frustrations on social media. As it turns out, other bank customers are experiencing the same problem, and by 5 a.m. the issue has evolved into a colossal PR problem. The bank must go into crisis communications mode before anyone has had their first cup of coffee.
A similar experience happened at a major bank, and to their credit, they were able to acknowledge the problem and correct it relatively quickly. Nonetheless, these scenarios illustrate that while mobile has opened up innumerable opportunities for companies to engage with and market to their customers, it has also raised tech performance expectations to an exceedingly high level.
Transaction Volumes Are of the Mind Boggling Variety
According to an article written by Peter Siddell at the IBM Hursley laboratory in the United Kingdom, titled “5 Things to Know About Why System z is Ideal for Mobile Apps,” there are some amazing mainframe transaction facts. For example, every second of the day across the world, 7,000 tweets are sent, 60,000 Google searches are conducted and 30,000 Facebook likes are made. But these numbers are paltry in comparison to the 1.15 million CICS transactions executed each second on System z—roughly 10 billion CICS transactions per day.
What do these metrics teach us? The number of customer-facing application users are generating an increasing number of mainframe transactions. Where there used to be a relatively small number of “internal users,” today there are millions of individuals interfacing with mainframes as they shop online, perform financial transactions or check the delivery status of a parcel. And while there are more people accessing the mainframe than ever before, it’s the portability of our connected devices that’s largely behind the increase in mainframe transactions. Imagine you’re traveling through Michigan and you decide to order something from an online retailer. Your retail order will kick off numerous mainframe transactions, beginning with your cell phone accessing a cell tower. As you search product inventory, make a purchase by entering your credit card information, receive authorization from your credit card company or bank and specify shipment instructions through a delivery service, you will hit numerous mainframes across several vendor IT environments. It is not just e-commerce that can generate multiple mainframe transactions. A simple task from a web browser, such as logging in to see your bank account balance, can kick off as many mainframe transactions.
With all of these transactions occurring on any given day—and more during peak holiday shopping times—it goes without saying that high customer expectations are putting increasing pressure on the mainframe to perform. A performance bottleneck in any of the previous scenarios sends costs skyrocketing as systems slow and teams work through problems. Revenue opportunities are put in jeopardy when consumers have poor application experiences and vent their frustration on social media. What’s more, the potential for transaction bottlenecks anywhere along the application delivery chain is further increased by a complex application architecture that transactions must traverse to even get to the mainframe.
Application Architecture Is Reaching New Levels of Complexity
With the mainframe as the cornerstone of the modern-day application delivery chain, enterprise application complexity continues to rise as application components span multiple platforms and operating environments. One illustration of how this complexity must be navigated is when an application must be tuned or modified. The changes have to be tested across a vast system of connected distributed and mainframe tiers. Yet, few IT professionals have a deep enough knowledge of all the components and how they interconnect to understand how a “simple” change to a multi-tier application can impact what’s going on downstream. Typically, staff has mainframe or distributed technology skills—but only a select few have deep understanding of the “end-to-end” path a transaction takes from a mobile device to the mainframe. These technology silos mean that when an upgrade or change is required, there must be collaboration and coordination among the teams to effectively roll out the changes. Importantly, those changes often have to be implemented across time zones and continents, further complicating the process.
Enterprise application complexity also rears its ugly head when application performance problems and exceptions occur anywhere along the transaction path. Without proper tooling that can span multiple domains—and is accessible to multiple teams across the enterprise—performance problems can be very difficult to diagnose and quickly resolve. Distributed and mainframe teams can spend hours, days or even months in the war room, pointing fingers and asking repeatedly, “what exactly is happening?” and “whose fault is it?”