X
Tech

How to simplify environments and cut TCO

With Web services, databases, multiple applications, and vendors all piled on top of legacy systems, it’s no wonder CIOs are looking for ways increase efficiencies and save money. Many IT departments have systems so bloated and complex that they’re crushi
Written by David Southgate, Contributor

With Web services, databases, multiple applications, and vendors all piled on top of legacy systems, it’s no wonder CIOs are looking for ways increase efficiencies and save money. Many IT departments have systems so bloated and complex that they’re crushing staff and fiscal resources.

But with more sophisticated computers and increased processing capabilities comes the possibility of reducing the server head count. Executives can boost their simplification efforts by standardizing things like operating systems and software and trimming back the number of vendors that support the enterprise. Simplification efforts can ultimately save money and improve service to end users—good news for everyone.

To learn more about simplifying IT environments, we spoke to Ken Pepple, IT architect in Sun Professional Services’ Asia Pacific practice and coauthor of Consolidation In the Data Center: Simplifying IT Environments To Reduce Total Cost of Ownership with David Hornby.

Pepple spoke on the topic last December as the keynote for a TeamQuest Corporation-sponsored User Group meeting. The meeting was held at the 28th International Conference for the Resource Management and Performance Evaluation of Enterprise Computing Systems organized by the Computer Measurement Group.

When not traveling in Asia, Pepple is based in the San Francisco Bay area, where TechRepublic caught up with him by telephone.

TechRepublic: Why have IT environments become so complex?
Pepple: From my point of view, we’ve moved into a new wave of computing. First Internet computing, and now we’ve started into Web services. In doing so, we haven’t removed many of the legacies in the IT environment. Instead, we’ve piled on more and more, making it complex. Because of that, people are starting to realize that we’re getting to the breaking point. Complexity isn’t just starting to affect costs and things like that—it’s also starting to affect the quality of service.

TechRepublic: What are the key concepts behind simplifying IT environments?
Pepple: There are three things, really: standardizing, centralizing, and virtualizing.

With standardizing, I talk to people about the different things they can do in their IT environments. They can start with the software and hardware in their environments. Beyond that, I talk to people about how they manage their IT environments, because that’s the way you can really simplify. One operational framework makes it easier to manage, easier to train people on, and easier to follow. Hopefully, that gives you higher quality of service and at a lower cost.

In centralizing, I talk about consolidating, providing fewer things for you to manage. A lot of the customers out there have 700, 800, 1,500 machines, 15 sites, and all these different vendors. For all of these people, I say, because of the advances in processing power, it’s now possible to have one machine do the processing of several machines. I’ll see anywhere from a 10-to-1 or a 30- or 40-to-1 [reduction in servers when people consolidate].

I’m seeing it all across the enterprise—people getting rid of mainframes, even their PC servers. Also in storage, direct attached storage-to-storage area networks or NAS (network attached storage).

There are a variety reasons they’re doing this—to lower the total cost of ownership, making it simpler to manage. Also, I see people doing it to leverage a new strategy. With initiatives such as disaster recovery scenarios, I see people saying, “I don’t know how to recover from a disaster recovery scenario with 1,500 machines. But if I have 10, I think I could recover from that.”

Virtualizing is the newest of the strategies for simplifying the IT environment. This will hopefully allow you to maintain more of what you want to manage—that is, applications—and keep you out of maintaining things that you don’t want to manage, such as hardware. Virtualizing lets you manage many hardware pieces as one piece, letting you start to automate some of this management.

TechRepublic: Are there things CIOs can initiate right now that won’t be costly in terms of dollars or time?
Pepple: The things I see people doing today can be done at any time. Things like cutting down the number of vendors and the number of applications you support can have impact on the total cost of ownership. So can simplifying the environment, making it easier for the staff to manage.

We see standardizing, though, as the first step in simplifying. A lot of clients do that before they move on to consolidating.

TechRepublic: For larger initiatives, what points should CIOs make to their executive boards to obtain funding approval?
Pepple: A lot of what we talk about in our book talks about how you justify a consolidation effort to your CFO or your CEO. We suggest people focus on reducing the total cost of ownership and proving you have a good ROI.

TechRepublic: How do you go about proving that?
Pepple: We try to generate what the total cost of ownership is today and try to project what the total cost of ownership of the consolidated environment will look like.

There are a variety of software packages that will help you estimate the total cost of ownership. Or you can simply go in and look at all your costs today for things like hardware depreciation, maintenance costs, staffing, environment costs, and software licenses.

TechRepublic: When creating a plan to simplify, what factors should CIOs consider in terms of operational hazards or liabilities? Pepple: A lot of times what you’ll look at on the risk side is two sets of risks. The first are political risks, the second are technical risks. With the political risks, obviously you’re taking different business units’ servers and putting multiple business units on the same servers. That can cause discomfort for the business units. The other thing to think about is internal finance—how you charge back to the business units. There are a variety of things you’ll need to look at.

On the technical side, you’re going to have a smaller number of servers, so you need to make sure that the new servers will still provide the same quality of service—availability, performance, manageability—the standard things you look at for your IT environment.

There are logistical risks to consolidating, moving from one vendor or one software to another. Obviously, you need expertise to be able to make that migration, as well as appropriate staffing that you’ll need to dedicate to that.

TechRepublic: What about issues in terms of personnel or staffing?
Pepple: The biggest thing we’ve found on either standardizing or consolidating is the organization of the team and who it reports to in the business units and the executives, and the amount of staff that’s available. It’s very key that you have a well-defined project structure and a well-defined means of communication back to the business units and the project sponsors—something that a program office provides.

TechRepublic: How much can it cost to simplify?
Pepple: It really depends on complexity and scale. You’re going to have some upfront costs to simplification. And hopefully you’re going to have a payback in four to five years that’s more than what it cost you.

Usually what we’ll find is that people consolidate and centralize a fair amount of storage onto a new server. The cost of maintaining the new service is significantly lower than the old environments.

For example, if I have 100 servers and it’s costing me $5 million a year in service contracts and staff and depreciation, and if I consolidate down to one or two machines that are only costing me $2 million a year, I’ll be saving $3 million a year over the next few years, whatever my depreciation cycle is.

Most of our clients look at payback or return on investment rather than upfront costs because all of them will be capitalizing this or even leasing, possibly.

TechRepublic: What personnel expertise might be needed?
Pepple: Usually I see an investment upfront because oftentimes you’ll need some expertise you don’t have. Often, there’ll be consulting or something upfront that you’ll need to pay for. But on the back end, you’ll either be able to reduce your staff or not hire people that you would have had to hire if you didn’t do this.

TechRepublic: Where can the process trip people up?
Pepple: I usually find that on most of the consolidation projects, if you’re going to run into problems, it’s in the initial planning and the expectations. I see that consolidation is pitched as one thing and it doesn’t come through as that in the end. So, you need to be clear about what expectation you’re setting when you’re going to standardize and consolidate. A lot of executives say, “I’ve heard about consolidation, it’s going to save me a lot of money.” But it’s very important to have very clear metrics and goals and objectives going in so you can communicate clearly back to the executives.

TechRepublic: What other advice do you have for CIOs who might want to tackle this?
Pepple: The one thing I see a lot of people do is an amazing amount of analysis on this. It gets put off because people look for reasons they can’t do this, rather than why they can. It eventually hurts them in terms of service.

I tell people to make sure they consider the opportunity costs and factor that into the decision. Also, CIOs should look for low-hanging fruit—easy projects to start with—and go from there

Editorial standards