“The shared services goal was essentially to say in order to for us to be effective as a company, we need to have information instantaneously from all of the places that are producing information―from all of the business activities that are occurring―you need to have access to that information very quickly and, from that, we wanted to put automated decision making criteria around a lot of it.”
Today, NetJets runs on UNIX for mission critical, high capacity apps and Microsoft for the desktop. They use Microsoft’s Development Suite for developing in-house apps like collaboration tools for decision making. All the information ends up in an data warehouse and he uses expert systems to make automated decisions around scheduling. The front end is Windows Presentation Foundation, the mid-tier is built in Java, and everything runs on Oracle databases.
“You can really boil all this stuff down to ‘What are we trying to do?’,” said Cullop. “We are trying to take the defects out of the process. Any places there are slowdowns or opportunities for defects to be introduced, you’re trying to remove those from that process. We got very, very good at this in manufacturing; where you have all that analysis trying to figure out how do I produce a million cans with no defects? The same principles apply across service processes as well.”
This is just one part of a multi-faceted philosophy that drives his decision making. From the technologies he’s going to field, to how IT is structured, to the role of the business in IT, Cullop’s main goal is to enable the to business to do what it does well, while keeping those things that IT does well inside of IT. In other words, let the business make business decisions―what projects get done, what (not how) IT can do for the business, ROI determinations, financials, etc.―and let IT make IT decisions―platforms, technologies, architectures, best practices, etc.
Whatever your core business proposition―whether its goods or services―you need to look at what are the core operational processes; find the dependencies and commonalities between business units from a high-level-flow perspective; and link all of the systems and processes together so you can automate them.
“That’s why I’ve always been a proponent of a centralized infrastructure strategy as it pertains to IT,” said Cullop. “There is the tactically ‘What do you do with the technology pieces?’ but first you’ve got get buy-in from the business” on the benefits of having a common or shared platform. “You determine business what we do, we determine how it integrates and what it runs on on.”
Since Cullop took over as CIO, the company has grown 50% but they’ve also cut power usage by 50%. What used to be 500 servers are now 27 rack-cooled IBM blades. This is part of NetJets’ Green initiative―which Cullop also must take into account when making technology choices―that stretches across the company from fuel consumption to IT. (They are funding research at Princeton to develop a cleaner burning jet fuel, for example.)
They’ve also gone from 18 UNIX boxes down to five AIX boxes that are bigger and draw only a little more power than the old UNIX big iron, but are more efficient. They’ve also gone from five storage area network (SAN) units to one. They’ve improved performance while at the same time cutting costs. To date, there has been a 200% return on these investments in real dollars, not expected ROI.
Taking on Inertia
Since there are not a lot of software vendors writing apps for the aviation industry, another part of Cullop’s job is to ride herd over the development teams. To help save money, Cullop works with Indian outsources. This saves anywhere from 15%-25% on overall development costs. More important than cost savings, however, has been getting development and operations to work together.
When Cullop arrived on scene, development developed (12 releases a year without QA testing) and ops basically did their best to hold everything together. There was little communication between the two groups nor were there any incentives for this to change.