For some, the jury is still out on whether it is indeed better to outsource to someone who can do it better and cheaper. Companies have been known to take IT back in-house, after they've outsourced it. Do you think there is a strong business case for outsourcing for all companies, or just some?
I think there is always an example where it just doesn't work, and people bring it back inside. In Sun's history, we've brought one contract back inside. We had legitimate reasons for doing that. I think it's a question of time more than anything. In any industry, there are organizations that are more progressive or on the 'bleeding edge' of technology, or more anxious to take advantage of those things. But in the outsourcing industry, it is just a matter of time before these large service providers find the right business model for a large number of companies. So it's probably not the perfect solution for all companies today but over the next five years, it'll mature to a point where, especially for midsize organizations that are trying to attract IT talent, it just won't make sense to try to staff IT organizations when you have these large, stable and mature service providers that can do the work for you.

Tell me more about the IT function that you took back in-house.
Years ago, we outsourced our desktop support, i.e. the thin-client technology support to a company in the United States. At that time we didn't appreciate that we were one of those leading-edge companies, and our desktop model was very different from what our service provider had traditionally been familiar with. And, I think to defend them a little, and frankly for ourselves, we were just pushing it too hard, too fast. I think if we did it today, that particular service provider would be more apt to provide support, but I think we just threw something at them which they were uncomfortable with, and both sides decided after two years to bring it back inside, to stabilize the environment and wait till the industry became more familiar with it.

Are you contemplating outsourcing desktop support now?
It's always on the table. But at the moment, we don't have any active plans in place.

I'm surprised. One would think that Sun would handle something like desktop support on its own.
Yeah, there are pros and cons to handling it yourself. In our case, the thin-client being a very standardized environment, there is really no reason why a vendor couldn't do it for us. But in this particular case, it was just so unfamiliar to our service provider years ago.

One big concern is the lack of IT talent. Which IT skills are in short supply?
Sometimes I feel lucky that Sun is so large and has such a good reputation in the industry that people want to come work for us. So we don't have the same challenge as many companies do. Nonetheless, there is definitely a shortage--whether it is Java programmers or programmers in general who understand enterprise applications.

Many students coming out of college today understand consumer applications. The kids grow up on consumer electronics and consumer technologies so very few of them understand how to put together an application like ERP (enterprise resource planning). It's not so much the basic technical skills that is lacking, but it's the knowledge of the business process. How do businesses operate? How do you build an application to support the business processes?

The second big gap is in security. Security is just a booming industry. There is network security, information security, and so on. Again, Sun is blessed with having a strong security department but for midsize organizations that may not be able to pay the same salaries or attract people in a way that companies like Sun can, it's a big issue because security is not going away any time soon, and yet you have to have the skill sets in your staff. So you have to pay quite a premium to attract people to these small and midsize companies, or bring in service level contracts to manage that.

The third is a general area that we're continuing to develop within Sun, which is described loosely as program management. It's the expertise to manage very large-scale multinational programs to deliver applications on time and within budget. These skills are not typically taught in the colleges and not typically things that IT people are very good at, so companies again have to develop the skills themselves, train people, or hire consultants.

How is the Oracle implementation coming along?
It's coming along well, and it's the largest systems project that Sun has ever undertaken. This is a perfect example of program management which I just talked about. By July 2008, we'll replace about 700 business systems or applications with one--the Oracle platform. So you can imagine that the complexity of managing a program of that size requires very sophisticated program management methodologies and capabilities, and that again is not typical in most IT shops. But the program is going well, we are scheduled to go live with our first deployment on Jul. 23. So the clock has already begun to count down, and people are becoming a little nervous and that's typical in such large programs.

What's the first application that will go live in July?
The first set of applications that will go live in July are focused mainly on the global general ledger, which is the basis of everything else. And beyond that the majority of the first phase is on the services areas, so it is the applications that Sun uses to support our end customers and clients, things like contract management. The second major phase of the program that is currently scheduled to go live in October will focus on what we loosely define as order management. That covers things like order processing and providing a single invoice to our customers. And the third phase currently scheduled for July 2008, is everything else but the predominance of that is in applications like manufacturing and logistics.

What was the impetus for this massive project?
It was really two reasons. One is significant cost savings. As you can imagine, turning off 700 applications saves a lot of money, or probably close to 1,000 servers that you can turn off. But just as important is business simplication across the company. If you look at how Sun has evolved through acquisitions, such as the StorageTek acquisition, our business processes have become overly complicated. And because we decided to use a very standard vanilla Oracle implementation, our business processes will be simplified. So today we don't have an easy ability to generate a single invoice, and Oracle will certainly make it simpler and cost-effective.

What's the standard operating procedure for you when an acquisition is to take place?
The first step in any acquisition is always a due diligence phase by various groups including IT and order management, and it's really to identify an inventory of the current applications of the other company and map that to our application environment at Sun. Generally, our philosophy has been to integrate the new company into our system and not leave their legacy systems behind, and that's exactly what we're doing with the StorageTek systems.

As each of the phases of our ERP application development occurs, we'll turn off StorageTek's old systems. In the long run, it'll become simpler, safe and less expensive to consolidate and integrate. Once that is done, it comes back to program management. We have assigned a full-time dedicated program management team to oversee the end-of-life, or ramp-down, of old legacy systems and to transition to the new systems. It's more about working with the business and having good program management skills.

What are some of the lessons learnt in handling such projects?
I wish they taught us this in school. One of the major lessons is that it is not always about the technology. Again, IT people like to think of the world from an IT perspective. It's hard to tell IT people to put IT aside, to forget the technology, and to sit down and understand the business and SLA (service level agreement) requirements, the sales issues, the market issues and service issues. And then go away and think how we might be able to apply the technology.

The second, and it may sound a little silly, is very good communication skills. I can't tell you how many times we've had good technical people who have difficulty articulating their ideas to the business. We spend time in our leadership development program to provide training to people to show them how to deliver their ideas more effectively.

The third thing they don't teach in school is the financial management in all this. If you think of three legs on a stool, financial management is another area where most IT people don't have a good understanding of how to put together a budget, or how to mange and track depreciation, or how to calculate the financial impact of, say, compliance issues in America. They don't teach you areas like how to read financial statements, how to communicate and talk in the terms that the financial community understand.

So you're saying that all these things should be taught in school today?
I attended a meeting with one of the major universities in California recently, and they asked me what they were not teaching our students. I gave them the same answer, and I think they recognize financial management and program management are things that they are beginning to teach their students, at least in California, and I think about how better prepared these students will be when they join the workplace.

Eco-friendly is the latest industry buzz. Is it just corporate social responsibility or a political move?
A little bit of it is corporate responsibility, and certainly in the United States, it is a political move for people to be more eco-friendly. But for the selfish CIO, the angle they should be more concerned about is that most of the solutions that drive eco-friendliness are also very good for data centers. They require less physical server footprint and they consumer less energy, so they drive costs down. Many CIOs in the United States will say they are eco-friendly, but they are really going for it to save money.

How do you suggest CIOs approach this?
The first thing is to look within the data center. Most data centers are running 10 to 15 percent utilization in the servers, but if you apply some of the new technologies, you can get the average utilization up to 70 to 80 percent. That allows you to turn servers off and save money. The other angle that people look at is the application of new technology that run at substantially lower energy consumption rates, which have direct impact on the energy bill that most CIOs are paying.

A personal example is we have two data centers in Europe and United Kingdom. We recently applied new technology to the data center consolidation effort and we've managed to reduce the square footage we need in the data center by 70 percent, and reduced energy consumption by almost 50 percent. So that's a serious and real cost-saving, and at the same time, it's an eco-friendly story as well. It's a very compelling story for CIOs who are focused on cost.

Are there challenges to building an eco-friendly environment?
I don't think so. I also don't think that it's right for every organization to think they have to become eco-friendly quickly. I think that it's an evolution thing. It can be as simple as telling you to go buy eco-friendly servers for a data center, for example. How important is it for the CFO to balance processor speed with cost and energy consumption? At Sun, we're espousing new metrics to capture these various dimensions of processing capability. So we're trying to come up with a standard metrics to evaluate the relative eco-friendliness, if you will, of one processor over another. So I think it can be as simple as looking at your next procurement and considering for the first time, perhaps, energy as one decision criterion in your purchase. As you begin to look at energy consumption in the data center and as those decisions are made, over time, you become more and more eco-friendly, so it doesn't have to be an overnight sort of thing.

Do vendors today actively talk about energy consumption in their sales pitches to customers?
Very rarely. But more so lately, particularly at Sun because I think we've got a good story in that space. But just recently, companies are beginning to be asked questions so they come prepared to answer the question nowadays, and that's something that was never talked about. It was just something that wasn't that important. But if you believe Gartner and others--their prediction is that in the next five years, the energy cost of running the data center will be more than the equipment in the data center. When the energy cost is more than the cost of running the equipment, people care. They will start to demand every hardware provider to be able to indicate the cost.

 Previous 1 2 

	Back to CIO Vision Series

Post your message here