Budget For Infrastructure And Shared Systems, Say IBM Top Brass
October 28, 2013 Dan Burger
The more that IBM‘s enterprise computing business changes, the more it stays the same. Last week at the Enterprise 2013 conference, a newly constructed crossroads of Power Systems and mainframe technologies, the big dogs in IBM’s Systems and Technology Group preached unprecedented changes in IT while also emphasizing centralized computing, a strategy that it has always promoted as the best way for enterprises to operate. The audience seems tentative about the first, while being solidly behind the second.
Analytics, cloud, social, and mobile are all barreling down the road at the same time. Individually they are more than a handful to deal with. As a foursome, they are more than the average IBM i shop can handle. The economics of IT has always been important, but not everyone gets the same benefits from the same projects.
There are many ways to maximize data accessibility and remove the silos of information, according to Tom Rosamilia, the senior vice president in charge of Systems and Technology Group and a former general manager of the Power Systems and System z divisions. “Technology is the easy part compared to people who are the difficult part,” he said in his keynote presentation. “This kind of change requires all types of collaboration.”
Collaboration has been blocked on the technical side as well as the business side. Both have a lot to do with people problems. Integration of systems that combine networking, storage, and computing have technical issues that are often easier to solve than the personal biases that prevent technical progress. The same type of biases and turf protectionism exist on the business side. Departmental fiefdoms in human resources, finance, procurement, and marketing to name a few have created their own islands.
C-level executives are sometimes oblivious to religious wars and political schisms and sometimes are participants. However, those who recognize this as an impediment to cost reduction and efficiency gains and then do something about it are the ones most likely to make progress happen. Shared resources, converged systems, and the infrastructure built to handle this is IBM’s vision.
“There’s a historic shift in the new age of computing, and infrastructure matters,” Rosamilia said. “We are moving from monolithic applications to dynamic services, from programmed systems to learning systems, and from proprietary systems to open systems.”
This isn’t a new message. IBM has delivered it for years. However, technology shifts that affect business rarely happen quickly, particularly in relation to the hype that leads to unrealized expectations. Then there are the cultural shifts that accompany the techno shifts. They also apply the brakes.
Rosamilia noted IBM’s investments in technologies such as Linux, open APIs, support for the OpenPower consortium, social business software, analytics, and flash storage are all indicators that Big Blue is prepared for the future of enterprise computing.
Any talk about the future of enterprise computing has to go through the reality of today’s IT environment. Although the budget belt has been loosened a notch or two, the grip on expenditures still feels like a dress shirt with a neck size one inch too small.
“The costs of IT have been rising,” acknowledged Steve Mills, general manager in charge of the combined Software and Systems Group. “But the main reasons are not the systems themselves. It’s the associated costs of operation–the server management, the admin costs, and the power consumption. There is better price/performance and better computing power. The cost of the physical infrastructure keeps going down and the cost of a unit of work is down.”
Most companies are spending close to 70 percent of their IT budgets on maintenance activities, according to widely circulated statistics. That leaves around 30 percent to be spent on innovation, which is not in short supply.
During his keynote presentation, Rosamilia said leading edge companies have reduced the amount of budget devoted to maintenance work to around 50 percent.
Controlling server sprawl is critical in managing IT budgets. This has worked out for quite a few IBM i shops that have consolidated remote boxes onto a central IBM i server. Some of those companies have converted those operational efficiencies into new projects such as mobile application development and increased security. Some companies just pocketed the savings.
“Most companies do not have an activity-based cost structure around IT,” Mills said. “They struggle to find costs and allocation is not easy to determine. Hidden costs are a factor. If you don’t do this, you will be making blind decisions. Some things look like values when all considerations are not taken into account. IT costs and the charge-back structure is traditionally a way to have projects for other departments paid for my IT. False economics get created and true costs of computing are masked.”
“IT budgets are cumulative,” Mills continued, pointing to efforts to instill some long-term planning in the minds of his listeners. “They include servers, storage, and networking–the infrastructure. There could be dozens, hundreds or thousands of servers and petabytes of data and the hundreds of thousands of programs that are run in any given environment.”
As an IT manager, it is important to understand what your company has relative to automation, optimization, and scheduling in a potentially integrated environment. There are significant investments in the sophistication of applications that may not be realized. And moving a unit of work in a sophisticated transaction processing system is not an easy thing to do, nor is it necessarily a wise thing to do.
A deeper understanding of what the system includes may abort a not-so-well-thought-out idea of jettisoning the system because it was underestimated. Migrations regularly involve additional scope than originally planned to match what was inherent in the undervalued system. Some end up in protracted dual environments, with cost and time overruns common.
Taking into account statistical multiplexing is almost always missed, Mills says. “You can load big systems in ways that can’t be done with small systems. Not understanding or not taking this into account can later add hidden costs.”
When comparing systems, optimization and utilization is all about finding applications that could be run together to save costs. Just like the process of eliminating servers, there is a process for the elimination of redundant software. These discovery stages identify how many apps are running and how many are necessary. It doesn’t take a genius to discover fewer apps mean lower cost. But you can add to that by noting that moving less data means using less server capacity. And better data control leads to improvements in processes and economic benefits.
The above formula can also be applied to improvements and efficiencies in shared storage practices.
Mills also advised that critical thinking about infrastructure can be applied to on-premise and off-premise computing. This includes platform as a service, infrastructure as a service, software as a service, and business processes as a service. The goal is simple: better business efficiency. There are companies providing services that are better at it than you are. But if you don’t understand what you are spending money on today, you might be disappointed at the amount of money you are saving (or you may not be saving any).
Although companies look to the cloud to save money, they are not analyzing how application integration is going to take place and the amount of data that will have to be moved back and forth. Activity-based costing is critical in getting this sorted out.
Prior to this year, Power Systems and IBM System z each had its own technical conference. Power Systems contains the IBM i, AIX, and Linux operating systems, while System z features z/OS and Linux. The amalgamation of these systems at Enterprise 2013 underscored IBM’s interest in riding data centers of the siloes of information that create limited access to data when and where it is needed. Easier said than done, right?