Unexpected Consequences and the Growth of Transactional Workloads
One of my college professors once rebuffed a comment I made in a sociology lecture about “unknown factors” with the statement, “the unexpected doesn’t happen very often.”
One of my college professors once rebuffed a comment I made in a sociology lecture about “unknown factors” with the statement, “the unexpected doesn’t happen very often.” I am sorry to say that, in my immaturity, I led the class in a chorus of derisive laughter. He was an engineer by training and I guess still had delusions of control.
In my experience, the “unexpected” happens far more frequently than we think it should and results in unintended consequences. This is mainly because of our limited ability to anticipate the future behavior of complex systems. And although something about “unexpected” usually implies “bad”, there are many cases of beneficial outcomes. For example, nearly everyone agrees that Tim Berners-Lee is the inventor of the World Wide Web, but did he really have any idea what his invention would turn into? Of course not — he was just trying to find a way to share documents between researchers!
And so it is with the growth of mainframe transactional workloads in the wake of explosive growth in mobile computing and ecommerce.
In the US, e-commerce transactions have been growing steadily since 2004 and in 2013 represented 6% of all sales, or $263B, an increase of 17% over 2012. I have no doubt that this pattern is the same across all the developed economies. Initially driven by a conventional browser and website combination, the new growth is now coming from mobile devices and includes app-driven traffic.
What would any of this have to do with mainframes? The answer: “a lot” — more than most people realize.
Mainframes are the hub of our worldwide economy:
If you use a credit card to make a purchase on a website it probably results in a transaction on a mainframe somewhere. Actually, it probably results in multiple transactions on multiple mainframes given the inter-bank clearing processes.
All of this growth in e-commerce has caused an unanticipated growth in mainframe transactions and workload. In the 1990’s, as many architects moved front-end applications off the mainframes on to client-server distributed platforms, the mainframe still provided the back-end transaction processing. Now these apps are being re-written again as web apps, mobile apps that tend to use canned services with RESTful APIs and designs that end up driving even more transaction traffic.
Looking forward, a significant amount of new transaction growth will be from new economies like China where nearly $1.6B mobile transactions were executed in 2013. Interestingly, these new BRIC (Brazil, Russia, India, and China) banks favor IBM’s mainframe technologies as they emulate the infrastructure of the established economies.
The world’s economy effectively runs on mainframes. That’s a good thing, because mainframes just work. Think about it: When was the last time you went to get money from an ATM and were told that the server was down? Despite the odd processing hiccup, usually caused by human failure, mainframes are incredibly reliable, secure and always kept physically secure as a priority.
As our global computing infrastructure becomes increasingly cloud-based it returns to an inherently more centralized model of computing, with fewer, larger datacenters. The mainframe has a growing role within this model as the platform of choice for fast, secure, cost-effective reliability. The combination of powerful mobile client devices and a reliable mainframe datacenter will take you into a secure, solid future – with as few unintended consequences as reality can manage!