Getting data centers to reduce power consumption has propelled major information technology vendors to join forces to attack this industry-wide issue. After all, no one company has the expertise to solve this complex problem alone. In February 2007, AMD, American Power Conversion, Hewlett-Packard, Intel, Sun Microsystems, VMware, and others came together to form The Green Grid, a not-for-profit industry consortium to promote energy efficiency in the data center. The organization presently has more than 100 member companies, which together provide all of the hardware, software, computing platforms, and components found in a data center.

Recently, Enterpriseleadership.org sat down with two of The Green Grid board members -- Larry J. Lamers, who oversees industry standards at VMware, and Roger Tipley, director of the The Green Grid at Hewlett-Packard Company -- to learn what this organization is working on. [Note: Their responses reflect the view of The Green Grid, not view of the board members' respective companies.]

EL: Why should CIOs be interested in your organization?

LT: We're trying to create an environment where power consumption and data center efficiency can be monitored in real time. We're also providing data and guidance to enable them to make better decisions about the design and planning of their data centers before they're built. We want to help data center managers make better decisions about equipment deployment and ongoing data center operations. If you plug in a new piece of equipment, how do you assess its energy efficiency before you buy it, and then after you buy it? That's our challenge.

EL: How do you collaborate as a group?

LT: We have a broad range of activities going on, including four active workgroups in the technical committee. And each of those groups has between three, and four taskforces working on specific problems. We have regularly scheduled technical meetings, which could be face-to-face or via teleconferencing; we use a variety of tools to support our Web site, and to manage our meetings, documents, minutes, and attendance. A third-party management company oversees the workgroups and committees to make sure they get everything done properly.

EL: What are some of these taskforces that comprise the technical  committee, and what are some of their activities?

LT: Our workgroup on data collection and analysis gathers and aggregates useful metrics about current data centers, such as performance. We have the problem of trying to understand how today's data center works -- that's why we're building a database of data center characteristics. Meanwhile, the operations workgroup looks at best practices, such as how systems should run for daily data center efficiency. This workgroup will define standards that support findings coming out of the other committees. Although each workgroup has a charter, all the workgroups interact with each other. They are collectively working on how to measure and to quantify data center efficiency. At this point, the technical committee and the workgroups have a good idea of what constitutes data center efficiency. They now, however, need to get some metrics in place so data center managers can gauge where they are, and what types of improvements they need to make, if any.

The technology and strategy working group looks at emerging technologies. One of this group's taskforces is working on what standards will be useful for building the data center of the future.

EL: What best practices have surfaced so far?

LT: We're looking at best practices differently than most IT organizations would. We're focused on being able to provide a way to measure the energy efficiency in a data center. We want to find a way to measure one solution at a time.

In March 2007, we published a common-sense way to measure the  efficiency of the data center. Some of these best practices included checking your floor files, checking the power configuration on your servers, and making sure your floor layout follows the guidelines that were set up when you designed your data center.

Often, you'll design your data center one way, but over time, you tend to modify things. Because of that, your systems aren't running optimally any more. You'll find that the highest performance per watt you can get out of a server is when it is highly utilized. Thus, you might think of moving to virtualization.

EL: What standards are you working on?

LT: Many industry organizations are working on specifications and standards. On the other hand, we provide what a lot of IT people need -- a reference guide or a checklist of what they should be doing. The standards should come out around the metrics and the measurement technologies. We might see some standards for how to accumulate or to assess what goes into the data center so you can manage it effectively.

EL: What are some of your members doing to make their products more  energy efficient?

LT: It depends on who you're talking to. Given that the work per watt of power consumed is going up, server vendors have made their processors more efficient than they were a few years ago. Vendors of racks, UPSs, and cooling equipment are improving the efficiency of those pieces. We're looking at how each piece is optimized and factoring them into the data center as a hole.

EL: You've just mentioned what the hardware vendors are doing. Well,  how are companies making their software more "green"?

LT: Both Microsoft and VMware realize that power consumed for a server at rest at low utilization isn't significantly less than the power consumed at full utilization. So these companies now provide software aimed at increasing the utilization of servers so they are up in the 80-plus-percent workload range. If you turn off these servers, they are running at 20 or 25 percent. And, if you take three servers that are running at 25-percent utilization and you put that workload at 80-percent utilization, you'll get some energy savings.

EL: What will the data center of the future look like?

LT: Future data centers will offer a more scalable infrastructure. If your workload goes up and down, you need to be able to adjust the power utilization. We've seen some data centers where, if you turn off all of the servers, you'd still not have zero power utilized in the data center. The scalability of your load has to match the scalability of your infrastructure.

The pieces that go into building a data center will be more tightly coupled during the design phase of future data centers. In the past, you hired disparate vendors to provide various components, such as power and cooling, racks, servers, storage, and wiring. In the future, people will want to know ahead of time that things will work together. These data centers will take advantage of their environment, as well as the prevailing, natural effects that are useful for power and cooling.

EL: What government agencies are you working with and how are you dealing with the differences in foreign data centers' power requirements?

LT: We have some alliances with various European commissions to look at the requirements of foreign data centers. Here, we'll try to localize some of our activities. We recently announced a collaborative relationship with the U.S. Department of Energy and Energy Star.

For more information about The Green Grid, go to www.thegreengrid.org.

--

Elizabeth M. Ferrarini is a writer from Boston,  Massachusetts. Reach her at elizabethferrarini@yahoo.com.

"Never compromise integrity" has become a long-standing principle of the $44 billion Chubb Group of Insurance Companies. Chubb offers both commercial lines and personal lines of insurance through its 120 offices in 29 countries.

Everything about Chubb propelled June E. Drewry, a 30-year IT veteran, to come out of retirement in 2005 and become global chief information officer of Chubb & Son, the organization that oversees the management of the Chubb Group of Insurance Companies. She has held IT leadership positions, including CIO, at Aetna Inc., Freddie Mac Corporation, and the Aon Group. She also has received numerous IT leadership awards from organizations, such as Women in Technology and SIM.

While Drewry's predecessor did a great job running the IT organization, she, nevertheless, faced several challenges: improve the IT governance model, drive innovation, and make sure her successor could fill her shoes. Enterpriseleadership.org recently spoke with Drewry to discuss how she's carrying out these tasks, along with her involvement in IT organizations and what she learned from September 11, 2001. Here's what she had to say:

EL: What's the leadership structure of your IT  organization?

JD: Our organizational model mimics Chubb's decentralized federated model, which consists of three business units. We have 1,500 IT employees who are managed by three geographical (or zone) CIOs, three business unit CIOs, one CIO for finance, and another CIO for claims. Each CIO, naturally, has a group of direct reports. A shared service manages our IT infrastructure for the U.S. Collectively, these CIOs comprise my senior IT leadership team.

EL: Can you describe how IT handles business process  improvements?

JD: We don't have a formal program for it. Our applications folks are embedded in our business units, and our IT federated model mimics the company's federated model. We do have centralized core resources that look at the leverage of things across business units, such as using a significant process or eliminating processes or workflow. If we have a particular benefit in one area, we then use the core staff to make sure we're exploiting it across the business units.

EL: After you joined Chubb, the company went to a more formal  governance model for IT. Why did you need to do this?

JD: I got used to joining companies where the relationship between IT and the business units needed improvement. Chubb presented a different scenario: IT had good alignment with the business units. Our work quality was high; projects got done on time and on budget.

My first task as CIO included getting all my CIO leaders together and finding out what our next challenges were. We agreed to work on enterprise-wide efforts. Some of our work had become captive in the business units. If we were going to get a real payback from IT, we needed to develop our enterprise architecture. It had to enhance our ability to leverage without having to re-platform. We also wanted to bring the business unit leaders together to do their strategic planning and to leverage technology. Specifically, we didn't want to find out individually what each business unit was trying to accomplish and then go off and do it in the background. This new process enabled us to bring IT governance and business governance together in the same room.

EL: How does your governance model work?

JD: We meet more often than IT organizations in other companies. In fact, my peers in other companies say it's unusual to meet as often as we do. Because it was so hard to plan, in 2006, the senior IT leadership team started to meet once every week for several hours. We meet every two weeks for a couple of hours more. Then, we meet off site once a month for two-and-a-half days. During this time, we go through the formality of governance where we review technologies and begin to push them further along. We're not allowed to discuss the day-to-day status of upgrades. That's a waste of our energy when we're off site.

EL: How do you keep track of projects for all of these  meetings?

JD: We have a dashboard. It's not a balanced scorecard. Some of our business units have their own balanced scorecard where they look at people, process, and technology. We look at a set of projects we've initiated as a group. We have a subcommittee that works on this weekly to keep things moving between those two-and-a-half days.

EL: Anything else you can say about your governance  model?

JD: Based on all of our industry reading and research, we learned that the IT governance process was the first step for getting things done right. So, if this was the case, we decided to run our service-level-agreement efforts through the governance process, along with our enterprise architecture and our creation of centers of excellence.

EL: Do you use any formal quality best practices, such as Six Sigma,  in IT at Chubb?

JD: Within IT we're not working on Six Sigma. Instead, we've identified those best practices used by each business unit or by the IT groups that support those business units. Before we can standardize on these best practices, we need to assign someone to make all of the business units aware of them. To this end, we can create a center of excellence to accomplish this task.

For external best practices, we belong to several networking groups where we have a different person represented at each one. This approach has created a broader network for us and more development opportunities for each group representative.

EL: Can you go into more detail about your center-of-excellence  concept?

JD: We've mandated functions such as risk management or payroll. Our shared services organization runs specific functions on behalf of the business units. They can't go outside the organization for these services. That's how we derived economies of scale in the U.S.

Then we have centers of excellence where we create areas that reside in a business unit or in corporate. For example, a center of excellence might support a pilot project in a business unit. We know there is a payback, but in the meantime, the group needs to educate others. We also have mature centers of excellence that set policy for the way we do things.

If what a center does becomes embedded in the way we do business, then we eliminate the center. Right now, we're looking at how we account for where all of the funding for centers of excellence should reside, even if the center is being managed by the business unit. Our goal is to make sure we capture good metrics on what business units are spending on these centers.

EL: Is the center of excellence concept the way you drive  innovation?

JD: Yes. We make sure that we don't innovate on the same issues in each business unit. We try to innovate once and replicate after that. Some people fail to see that. If we see a real payback from something a business unit has done, we need to have an evangelist get the message out to the other business units. Another part of the evangelist's job is to find the business cases and to get in front of the strategic business units. It's up to them to make good business decisions, now, or a year from now.

EL: How are you going about grooming your successor?

JD: That's easier said then done. Long-time service has become a hallmark for Chubb. My predecessor was CIO for 16 years. This company understands that people are its greatest asset, and treats them accordingly. Most companies pay lip service to this.

I came out of retirement to work at Chubb with the condition that I'd be here three to five years. During this time, I'd be grooming my successor. My predecessor had a pool of excellent candidates who were considered for this job, and I've added other candidates to this pool. This is the first time in my career where my primary responsibilities have been to find my successor and to carry out all of the day-to-day CIO responsibilities.

EL: So what have you learned from finding your  successor?

JD: Looking for a successor has been an eye-opening experience. I've totally change the way I work; I wish I had done this 10 years ago. Most IT executives spend more time doing and less time grooming. Now, I delegate probably five times more than I ever did in my life. My job is to give my prospective successors enterprise leadership experience, which differs from what either a strategic business unit CIO has or a zone CIO gets.

EL: Do you think the CIO role should be rotational?

JD: It becomes a way to resolve issues with the business alignment of IT, or with business units that want to be involved with technology. Chubb is the first company I've worked at where business units had great alignment with IT. Business people here truly understand the value of technology and deriving more value out of it. So, why would we want to rotate a finance person who understands the value of technology and who accepts us as a perfect partner?

On the other hand, IT isn't going to set up financial policies, but we understand them, along with understanding underwriting and our markets. If you get the depth of knowledge right in the IT organization, then I don't know how much rotation buys you.

EL: You've participated in a number of IT venues, such as the CIO Executive Summit, and have been on the board of SIM. Why have these organizations been important to your IT career?

JD: The network has been invaluable to me. That's something I've been trying to teach our younger IT staff. All of the book knowledge in the world can't help you to deal with the reality of executing and getting things done.

The more I've move from one company to another, I've realized that the same issues exist everywhere; only the names and faces change. Organizations, such as SIM, provide a safe place for getting valuable input from IT professionals. You can use their input to resolve problems or to step back and make a decision. A lot of growth comes from this communication process.

The insurance industry, like many other industries, has its own group of organizations, such as ACORD. Because many industries are so competitive, one can no longer feel comfortable talking to peers at an industry venue. SIM, for example, provides you with a cross-industry network where you can talk about, say, your service-level agreements.

As for speaking venues, such as the CIO Executive Summit, they have been great places to make connections and to source IT candidates.

EL: As CIO of Aon, you had to deal with the tragic events of  September 11. What did you learn from this experience?

JD: I learned that it's all about people, not IT, when something like that happens. Yes, it's great to be able to transfer the work to another data center, which we did. That day, Aon lost 175 people. The entire company focused on trying to find them. The people who survived were in shock. Most people weren't psychologically prepared to go back to work. The last thing we wanted to do was to pester them about picking up a laptop so they could work from home.

--

Elizabeth M. Ferrarini is a writer from Boston,  Massachusetts. Reach her at elizabethferrarini@yahoo.com.

How are you getting ready for all of the disruptive technologies that will fuel the "telecosm"? If you're unfamiliar with the concept of the telecosm, perhaps you need to join the "Gilder Age." Coined by bestselling author, economist, and technology guru George Gilder, the telecosm is a world with an abundance of bandwidth and a scarcity of transistors, due to the ever-decreasing size of handheld devices.

Perhaps the best way to get to know this new world is to attend the annual  two-day Gilder Telecosm Forum venue, sponsored by Gilder  Publishing and Forbes. There, you'll find hundreds of CEOs, technology entrepreneurs, engineers, academics, investors, and money managers coming together to learn about all of the technologies that will make the telecosm possible. Many of these companies have appeared in the Gilder Technology  Report.

Of course, Gilder's tech visions might not emerge next week or even next year, but he always seems to be right about predicting what could happen a decade or two down the road. He's the bestselling author of Wealth  and Poverty. And, his collection of tech and business books include  Recapturing  the Spirit of the Enterprise, Microcosm,  Life  After Television, Telecosm,  and the Silicon  Eye. A frequent contributor to Forbes and The Economist, Gilder pioneered the formulation of supply-side economics when he served as chairman of the Lehrman Institute's Economic Roundtable.

Enterpriseleadership.org recently spoke with Gilder about the ins and outs of  the telecosm. Here's what he had to say:

EL: What's at the heart of the telecosm?

GG: The telecosm revolves around what we call the "exaflood." It's the overwhelming flood of 1018 data onto the Internet, caused by the increasing transformation of video form standard low definition, such as VGA, to increasingly high-definition forms. For example, when YouTube.com goes high definition, Internet traffic will rise 100 fold.

EL: How do you go about selecting the portfolio of companies you  report on and invite to your annual Telecosm venue?

GG: We work through my paradigm. For example, we've seen a 1,000-fold increase in video traffic on the Internet, along with an inexorable increase in video resolution. Once people go to high-definition television, they don't go back to low-definition images. As a result of these developments, we're seeing more emphasis on video worlds, such as MySpace.com and YouTube.com. Analog images or analog sounds are now being rendered and transmitted as IP packets and Ethernet frames. Ordinary Intel processors have not kept pace with the movement of this type of data across the Net.

The telecosm has to adapt when the exaflood of data produced each year hits. We try to figure out what companies can readily adapt to change. For example, graphics processors fit the bill. Other good candidates include companies that code and decode various image standards. A real-time processor has to be at least 30 frames a second or, better still, 60 frames a second to give a true immersive experience. Enter Sigma Designs.

The IP Ethernet transformation at 10 gigabits per second moving to 100 gigabits per second made us look at companies offering fiber-speed devices. We think EZChip has the best network processor. It hollows out the router by performing the router function in hardware similar to the way the Intel processor performs in the IBM PC. Other functions, such as security, also have to be performed at fiber speed. Raza MicroElectronics, Cavium Networks, and NetLogic Microsystems process packets  at wire speed.

EL: What innovative technologies can CIOs get their hands on to bring  the telecosm to their organization?

GG: Trusted platform  modules, a technology that has been around for a decade, are now being  incorporated into PCs. For example, little vault chips from Wave Systems sit on the edge of the network, performing a range of security functions, such as encryption and decryption. This development changes the entire structure of the network and the industry. You can have established authentication through various biometric schemes. With a trusted platform device, you also have absolutely secure authentication on the edge of the network. You don't need to have security devices and security processes attached to routers and other devices in the middle of the network. These devices actually create new opportunities for attack, rather than shielding machines from attack.

The Trusted Computing  Group, a not-for-profit standards organization for the trusted platform technology, comprises companies such as Hewlett Packard, IBM, and others that have adopted this technology.

EL: Where does optical networking fit into the telecosm?

GG: Most of the optical networking advances that began before the dot.com crash have been resurrected, and are making the network faster than the backplane of the computer. This transformation in computing has brought the demise of the wired LAN, the theme of Telecosm 2007.

Optics is one of the technologies where we've seen a 100-fold rise in network traffic just through that transformation. Optics is going to play a big role in transmitting photo-realistic images and video across the Internet. The capability of optics has been expanding three times faster than electronics for the past decade.

EL: To create the telecosm, what types of technologies are we seeing  with networks and network devices?

GG: Meanwhile, storage and bandwidth, which we call storewidth, has also been expanding at this pace. This development makes it possible for a lot of devices to connect to 10-gigabit networks, a move from the current Ethernet gigabit. All the electronics have to run at fiber speed, not finger speed.

Across the network, these devices have to convert from 10-gigabit streams into electronic processing, pass them over, route them, switch them, and classify them. All of the various functions that have to be applied to packets across the network need to run at a very fast speed. They require new architectures and new kinds of devices. Those are the companies we're looking at. Luxtera handles all of the optical  functions, short of the laser on CMOS microchips.

The network processor we focus on can perform millions of steps on packets coming in at 10 gigabits a second. We look at all of chips, devices, systems, and computer architectures, and new materials needed to couple all of the devices on the network edge.

EL: With the telecosm, you reduce the number of transistors. Can you  talk more about this?

GG: Terry Turpin, an optical engineer who has spoken at Telecosm about the fibersphere, has found that you can have as many as 14,000 of these wavelengths on a fiber without deterioration of the performance. The performance of fiber improves with the multiplication of wavelength paths down the fiber. This phenomenon of optics repeats the Moore's Law phenomenon in silicon. In other words, the more closely together transistors are on chip, the easier it is to produce faster, less expensive and better chips. That's the magic of microelectronics.

Meanwhile, microchips are moving to copper metalization from aluminum metalization. Because the transistors are closer together, you can accelerate the speed of the electronics three, to five fold. This transformation immediately affects the type of copper you can have. We focus on all the companies that are enabling that change.

EL: Are you working on another book?

GG: I'm 80 percent of the way through a book called  Analog. It's about two CalTech professors, Carver Mead and Richard  Feynman, who a taught a course on the physics of computation. Their body of work produced an entire series of books. Feynman eventually joined Thinking Machines to work on massively parallel computers. Carver Mead, the Gordon Moore Professor at CalTech and co-founder of Intel, has either launched or done design work  for more than 25 semiconductor companies. Mead researched and named Moore's  Law.

EL: Some CIOs say they don't know how they would use virtual worlds  such as Second Life. What would you say to them?

GG: A transformation in communication is underway on the Net. You're going to experience a new type of video conferencing and collaboration. You'll enter a virtual space where you'll be present in a physical-like way in that space. Here, you'll experience the richness of a face-to-face exchange, rather than through avatars. Everyone will be able to see you write on a virtual white board or to project PowerPoint slides. You'll even be able to run executable programs as desired within your space. This is where the technology is moving to.

EL: What are the key technologies that will enable these realistic  virtual worlds?

GG: One of the technologies we're stressing at Telecosm includes graphics processors. They're everywhere. The graphics processor in your cell phone uses the same ATI instruction set as a supercomputer. To this end, you have a  universal instruction set that's about as common as the X86 instruction set in graphics processors  These graphics processors will be able to create virtual worlds, such as Second  Life and the forthcoming CitySpace.

Using a 3D programming language called OTOY, and the instruction set of the instructor processor, Jules Urbach, in a matter of hours, was able to create photo-realistic figures at 30 frames per second for the Transformers movie. He would have taken days to do the same thing with some of today's technologies. With OTOY, the CPU hardly plays any role in the activity. You can use it with anything from a supercomputer to a cell phone, which can be scaled to accommodate a big screen immersive environment.

We have the gaming industry to thank for the development of the massively parallel graphics processor. This technology has moved three times faster than PC microprocessors. Think about it. AMD bought ATI. Intel has a project  underway that uses massively parallel graphics processors. Look at what Nvidia is doing.

--

Elizabeth M. Ferrarini is a free-lance technology and  business writer from Boston, Massachusetts. You can reach her at elizabethferrarini@yahoo.com.