Capacity planning and performance management should be priorities today. If you want to save money, know what you have on the floor and how it's performing.
Way back when "computer" effectively meant "mainframe", techies devoted their days to gleaning everything there was to know about the performance and capacity of their organisation's systems and applications. They researched their computers as diligently as a scientist researches a human genome today.
They could well afford to. With mainframes costing a cool $10 million or so and the techies' salary coming in at just $25,000 a year, any techie who, after six months spent tweaking the system, could announce they had delayed an upgrade or the acquisition of a new box for a year was a hero. By the end of the mainframe era service level agreements and measurements of response times were commonplace, and computers were running as efficiently as . . . well, computers.
Then, says Fujitsu senior business development manager Mike Tsykin, along came the mini computers, which the techies used to laugh at, and the micro computers, which they used to ignore. And the concept of transactions went out the window, and the implications of that for capacity management were both profound and largely overlooked.
"We can't measure transactions any more," Tsykin says. "We went from one single mainframe, which was supporting this wide variety of users, to a distributed infrastructure where each department effectively has its own network and small servers, and that supports the users. In other words, 'the network is the computer' - the Sun paradigm.
"Even IBM went that way, with the turn back to virtual machines and virtual computers in the same box, but it makes no difference really: they're all different, separate and they're all cheap, but they're not less complex. Windows is maybe not quite as complex as a mainframe system, but it's dammed complex enough. We can't spend anywhere near the same sort of amount of time on optimising and learning and doing things. And the rate of change is enormously accelerated so our knowledge gets obliterated; mine was obliterated perhaps six or seven times over in the last 20 years."
As a result, capacity management these days is just not done, Tsykin says, except in the cases of the largest systems. And that is a great embarrassment and shame, not to mention - ever since that distributed infrastructure went mission-critical - potentially damaging, because performance slowdowns caused by overtaxed systems could end up hurting companies big-time. Indeed Tsykin says when it comes to computing, organisations stand poised at the edge of a landslide. Distributed computing has got to the point where the infrastructure is becoming ridiculously cheap, and that infrastructure is not optimised. As organisations move towards blade computing models he says they increasingly need to adopt a distributed network operating environment, and that environment simply will not work effectively without capacity management.
"You've got to make decisions on some basis. Where are you going to run all of this? And that is where capacity management will come in, in the future," Tsykin says. "It's not too far away. But for now, what CIOs should be thinking about is quality of service. What is the quality of service they're delivering to their users? Are they actually supporting their users as they should be supported? Or are they doing too much? And it is possible to do too much quite easily. Are they supporting those who scream loudest and not those that return to the businesses? They don't know, but capacity management can help."
In a well-run IT shop, experts say, CIOs looking to ensure the stability of their environment should not only keep planning for variations in the use of the system - such as daily and seasonal changes in user and transaction levels - in the forefront of their mind, but also the need to ensure the system stays up and running at all times. Meta Group says the days when capacity/performance management (C/PM) were riven by vague corporate directions and wishful line-of-business thinking are slowly coming to an end, as besieged CIOs learn to "think" platform utilisation and eye consolidation benefits. The research company joins other research groups and analysts in warning the time has come for C/PM to seize the day and use strong processes and technical skills that can translate operational efficiencies and effectiveness into improved bottom-line performance.
"The key goal for C/PM is to be able to provide a holistic and measurable view of the IT environment and describe its relevance in business terms. This requires the proactive management of IT resources with direct input from application and infrastructure planning groups," Meta Group says. "As opposed to just monitoring resources, new-age capacity management ensures IT infrastructure matches the changing demands of the business in the most cost-effective (short- and long-term) and timely manner."
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.