IT Management

It has been 45 years since Intel’s co-founder Gordon Moore made the observation that came to be known as Moore’s Law. And, while the original observation was simply about the number of components in integrated circuits doubling every year, the spirit of this geometric growth in resource availability has pervaded IT ever since.

Of course, back in the early years, we always needed more than we had, so Moore’s Law couldn’t operate quickly enough. As a result, much care and effort were exerted to ensure that programs didn’t use any more resources (memory, CPU time, or other) than they absolutely had to.

Still, the demands of processing often kept up with or outstripped available resources, resulting in the development of a range of “computer laws” attributed to everyone from Murphy to Swiss computer scientist, Niklaus Wirth, including the famous and variously attributed, “Any given program will expand to fill all available memory.”

So, as business computers accrued more and more memory, external storage and speed, the processing requirements placed on them continued to drive demand for further enhancements.

About a decade into the life of the mainframe, an important set of requirements began to more clearly resolve: Reliability, Availability, and Serviceability (RAS). Close attention was paid to ensuring the mainframe wasn’t only big and fast, but strong as well in a way that met business requirements and justified the hardware, software, and personnel costs associated with it.

The insatiable desire for more continued as the PC revolution dawned, and soon it was possible to have your own private chunk of computing power on your desk, giving you access to the mainframe and a plethora of PC applications—from spreadsheets and word processing to games.

Now, while the speed and storage capacity of computers continued to escalate, three other things began to grow: the number of distributed computers, the number of people required to support them, and the complexity and variety of individual configurations.

Suddenly, despite reducing costs for a given amount of CPU capacity, memory and external storage, the costs of IT to the business began to creep upward. But it wasn’t all in one place;

rather, it grew like weeds in every part of an organization, as everyone wanted the status, and sometimes the functionality, that a PC brought.

So, in a scenario that brings to mind a number of different “Dr. Seuss” stories, personal computing grew so rapidly it swamped many organizations, as individual users became the masters of their own systems, but soon began to demand support from elsewhere.

Now here’s the funny part: In any given organization, one of the key factors in political power is the number of people in an area. So, the sudden arrival of a vast number of inexpensive, mildly functional PCs was followed by the arrival of a giant cohort of newly minted PC support people, most of whom had no background in the RAS disciplines the mainframers practiced. But there were many more of them than there were mainframers, which gave them a greater voice in the running of business IT.

If this were a “Dr. Seuss” story, at this point some eccentric hero would step in and put everything back on the mainframe that belonged there and replace all the PCs with graphical terminals attached to large, easily managed systems that weren’t user-breakable.

Guess what: That’s where we come in. I’m not insisting you have to be eccentric to be a mainframer (though it helps), but it’s time for each of us to take the heroic step of standing up and telling our colleagues and management that more isn’t always better: more costs, more complexity, more vulnerability, and ever more support staff requirements aren’t the kind of “more” that a good business IT environment should be striving for. 

It’s time to remind the world of business IT that the “more” they’ve always benefited from is more reliability, availability, security, stability, real affordability, and business value generated per IT person. After all, isn’t your business trying to control costs and do more with less? Isn’t IT looking to increase agility to support the business? 

Add a mainframe, and everything clicks.

Reg Harbeck works for CA Technologies. For more than two decades, he has worked with operating systems, networks, security, and applications across mainframes, UNIX, Linux and Windows, and traveled the world, presenting to IT management and technical audiences at numerous IBM, industry analyst and user group events.