We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.


SPOTLIGHT: Ron Powell & Glenn Keels of HP Discuss Cures for Aging Data Centers

Originally published August 24, 2011

BeyeNETWORK Spotlights focus on news, events and products in the business intelligence ecosystem that are poised to have a significant impact on the industry as a whole; on the enterprises that rely on business intelligence, analytics, performance management, data warehousing and/or data governance products to understand and act on the vital information that can be gleaned from their data; or on the providers of these mission-critical products.

Presented as a Q&A-style article, these interviews with leading voices in the industry including software vendors, end users and independent consultants are conducted by the BeyeNETWORK and present the behind-the-scene view that you won’t read in press releases.

This BeyeNETWORK spotlight features Ron Powell's interview with Glenn Keels, director of marketing for HP’s Hyperscale Business Unit as a part of the Industry Standard Servers and Software organization. He is responsible for lifecycle marketing of HP ProLiant servers, and the market’s broadest range of converged infrastructure products, services and solutions. In this interview, Ron and Glenn discuss the challenges that CIOs face today as they work to ensure they can meet the rapidly changing and growing infrastructure demands – and save money at the same time.

Glenn, your area of expertise at HP is Converged Infrastructure Product Services and Solutions. I thought we'd start the interview by looking at some of the factors that are currently impacting the infrastructure of today's enterprises. Our readers are no strangers to the data explosion. They're all experiencing it firsthand, especially from an analytics and business intelligence perspective, and many are searching for techniques and tools to help them manage this data efficiently. When you add to this complexity things like cloud computing and demand for mobile solutions, it seems like a lot for an aging infrastructure. What are your thoughts?

Glenn Keels: You're right on. Bottom line, we believe CIOs are challenged today more than ever. These are typically challenges of really being able to move from a manager of the infrastructure to being a more strategic part of the enterprise and transforming IT into a broker of services. The need to increase agility while reducing cost is omnipresent. Demands from the lines of business for more exacting and more precise service levels to be more competitive is increasing, and business cycles are becoming more and more unpredictable. Underneath all of this is a rigid and aging siloed data center infrastructure. The average age of today's data center is actually a whopping 14 years old. Go back to where you were 14 years ago. Can you imagine trying to predict your IT facility requirements back in 1997? It's almost impossible. By our calculations, new investments and needed data center capacity have been delayed around 37% between 2008 and 2010. This is due to the global economic uncertainty we've all faced. But as we're starting to come out of this recession, new data center capacity needs are expected to double by 2014. Data center capacity really is in crisis. But using traditional monolithic brick-and-mortar approaches to solve that data center capacity simply exacerbates that problem of rigid, aging siloed infrastructure. Here's why: It simply takes too long. The average time to plan, construct, and commission the average data center is two years or more. Not only does it take a long time, but it also implies that CIOs have a crystal ball and can predict enterprise needs, competitive moves, technology innovations, and economic cycles years in advance. This is a challenge for even the best CIO.

Existing monolithic approaches to data center capacity cost too much. A relatively small data center capacity with an IT load of 1.2 megawatts can cost tens of millions of dollars, and those are dollars that could have been used and redirected toward enterprise innovation, as opposed to keeping the lights on.

Lastly, brick-and-mortar data centers waste too much energy. In fact, the average brick-and-mortar data center wastes as much energy as is actually used by the devices that are performing the work – the servers, the storage, and networking. That energy costs money and is in short supply. Frankly, to be better stewards for our environment, enterprises should be reducing energy usage. In fact, by 2013, most governments are expected to mandate a basic level of data center energy efficiency. That’s less than two years away! At HP, we believe that our converged infrastructure portfolio – and, within that, our modular data center offerings – can fundamentally improve that situation. This approach allows CIOs to reclaim time and money spent just to keep the lights on and redirect that to innovation, thereby creating a much more powerful, agile, instant-on enterprise.

Well Glenn, you bring up a key consideration for many enterprises today – the need to limit the amount of energy that data centers use today. I know Europe has instituted some stricter standards and the area of carbon credits has become a key issue. Let’s now talk about converged infrastructures. Can you explain what you mean by a converged infrastructure and how that helps enterprises improve their agility?

Glenn Keels: Sure. Very simply, converged infrastructure is a blueprint for a data center that accelerates provisioning of IT services and applications. The key is integrating – in an open, industry-standard way – all servers, storage, networking, security, power, cooling, infrastructure and cloud services into shared pools of interoperable resources. And all those resources should be able to be managed through a common management platform. The net result is that a converged infrastructure lets organizations transform technology into a business advantage by being more adaptable and able to innovate rapidly as opposed to being brittle and rigid. And again, back to what I was describing earlier, this approach allows moving from a manager of infrastructure to a broker of services.

Most importantly, converged infrastructure that is based on open industry standards delivers the value that businesses are now requiring of CIOs: faster time-to-value in applications and revenue, reducing the cost of acquisition and implementation, more responsibly and quickly evolving to changing business requirements or changing economic cycles, and reducing risk along the way.

A key piece of that converged infrastructure is our Data Center Smart Grid, designed to extend the life of an existing data center and provide more modular, open industry-standard ways to expand data center capacity when required. We call that Data Center Smart Grid. And again, within that is our Performance Optimized Data Center offering or what we call PODs.

Now, you're bringing up one of the things that really is very exciting. In my prior career, I was involved in the manufacturing sector, and capacity management was a huge focus for that industry. You recently announced the HP EcoPOD, and one of the things that I read about it was that it uses 95% less facilities energy than a traditional brick-and-mortar data center. This sounds like a data center manager's dream come true for managing capacity. Can you tell us more about PODs and then tell us about the EcoPOD?

Glenn Keels: I’ll give a little bit of background, and then I'll go into the EcoPOD. About five years ago, we – as well as other vendors and customers – began leading what we call a revolution of approaching new or temporary data center capacity with what most people referred to back then as "containerized data centers." These containerized data centers provided energy savings to a certain extent and cost savings, and we could get this to customers faster than a traditional brick-and-mortar approach. We also could preconfigure it and fully load it with IT. We called those Performance Optimized Data Centers or HP PODs. About two years ago, we moved to next-generation PODs that still looked like containers but were actually very different. They still fit on a truck and they still had that shape of a container, but they were more modular, less costly and more energy efficient.

At that point, more and more types of customers, not just your usual suspects of extreme scale specialized environments, but more mainstream customers began to look at PODs as a true alternative to brick-and-mortar data centers. They began to adopt this new form of "data center economics" – economics of velocity, or speed, energy efficiency and then much more affordability. Just last winter, we took this to the next level by introducing POD-Works, the world's first assembly line for data centers, where we could assemble, configure, test, and deliver a POD fully configured with IT in weeks as opposed to years. In fact, just last fall we shipped 22 PODs to one of our inaugural POD-Works customers. We shipped 22 40-foot chilled water PODs (HP POD 40c) packed with more than 40,000 servers along with storage, and network, and other converged infrastructure – all preconfigured and tested in just nine weeks. This gains for this customer were huge in terms of speed, energy efficiency and affordability.

Today, we're proud to have customers around the world in broad industry adoption taking a look at and deploying PODs. It was a little more than a year ago when we started to see this move from early adopters to more mainstream adopters – crossing the chasm, I guess you would call it. We started a little bit further ahead. We started to listen to customers who you, again, wouldn't expect to take this modular approach, asking us for more. They saw the benefits of PODs and modular data centers as a whole but needed to make it a more natural, viable alternative to more markets, more industries, and more operational environments. Here's basically what they told us: They said this could be something big. In fact, they said bigger than you can imagine, HP. Ultimately, we cannot only decrease cost, increase agility, and save energy, but this can be the future – a time when IT and facilities are purchased together, deployed together, managed together, even retired together, all completely integrated. But you need to do more. You need to deliver even more CapEx savings. You need to deliver more energy savings. You need to do those two things while still being able to deliver it within weeks. But, most importantly, we need enterprise-class features, and we don’t want to sacrifice serviceability and some of those enterprise-class features when adopting a modular containerized approach.

So we started putting our top brains on it and realized that we could deliver all that was being asked. And if we did that, then we'd be able to accelerate this crossing the chasm and see more mainstream adoption of modular data centers.

Our answer, recently announced, is nicknamed HP EcoPOD. The formal name is the HP POD 240a, which is two 40-foot IT modules with adaptive cooling. It’s the first in a new category of PODs designed to deliver even better savings and accelerate this move to modular data centers. We've actually called it the world's most efficient data center, and it answers all those challenges that I just described. Let me go through a few of them.

First, we've taken a no-compromise approach to serviceability while increasing the density. In fact, one EcoPOD can equal the computing power of a 10,000-square-foot data center in one-tenth of the space. We've invented a new technology called HP Adaptive Cooling that introduces free air-cooling using direct expansion modules that intelligently adjust cooling based on IT load, policy, and climate.

Second, we've expanded the POD environmental control system to make it more comprehensive from a management perspective including being able to dial in that efficiency with HP Adaptive Cooling. When you compare this to brick-and-mortar facilities as an alternative, we really believe that it delivers that breakthrough value that's going to accelerate adoption. It can be completely configured and deployed in three months as opposed to 24 months – 88% faster. From a cost standpoint, it can be deployed for a quarter of the cost, 75% less. Most fascinating to me is that from an energy standpoint, it's darn near perfect with a PUE [power usage effectiveness] as low as 1.05. That can deliver 95% better facilities energy savings when you compare it to brick-and-mortar.

What does that mean? Let's take a look at each one of those savings. First, the deployment time. It should go without saying, 24 months versus 3 months. Again, it's not just that time you save, but you also begin to enjoy the benefits of right sizing and integrating IT planning with facilities planning. Again, by our estimates, it takes about two years to deploy the average brick-and-mortar facility, spanning multiple stages, multiple vendors, and multiple organizations to plan, as well as deploy, construct, and commission. Compare that with a HP POD that takes three months, again 88% faster. But equally important, I talked a little bit earlier about that crystal ball and how even the best CIOs have trouble predicting acquisitions, business requirements, technology needs, capacity needs – you name it – years in advance. By taking this modular pay-as-you-grow approach, CIOs can confidently plan their facilities requirements, knowing they won't be stuck with either not enough capacity or too much capacity as their enterprise needs change.

Let's look at cost. Beyond the benefit of that time, there are significant benefits from combining extreme densities, serviceability of same modular approach and adaptive cooling for as low a quarter of the cost when compared to brick-and-mortar. Let's look at an example: A traditional tier three, brick-and-mortar, 1.2-megawatt data center can cost as much $33 million to deploy, by industry estimates. With the EcoPOD, that cost can be as low as $8.3 million – that’s $25 million in savings in CapEx alone, all in one-tenth of the space.

Now let’s take a look at that near-perfect energy efficiency. Most data centers today have a power usage effectiveness or a PUE ratio between 2.0 and 2.4. That means that for every watt used by the devices that are performing the work – the servers, the storage devices, the networking, the switches, etc. – there is about 1 to 1.4 watts wasted in housing, powering and cooling that same IT equipment. By reducing that PUE to as low as 1.05, only 5% of additional energy is required to perform that same function. If you take that same 1.2-megawatt data center I just talked about and use ten cents per kilowatt on average and a PUE of 2.4 compared to an EcoPOD, that's going to amount to around $1.5 million in energy savings alone per year at a data center level. So with environmental mandates looming and enterprise sustainability being top of mind for CIOs, not only is this good for the enterprise bottom line, but also it's good for the environment.

I think one of the nice things about it too is that because the EcoPOD has such a small footprint, you could take an existing data center and actually increase the power just by substituting or integrating the EcoPOD within an existing data center. You could increase the capacity without having to increase the overall space taken up by the data center.

Glenn Keels: Exactly. It’s a pretty powerful solution.

That is great. One of the other things that you mentioned in your release was about cloud computing. How are you helping organizations in the area of cloud computing?

Glenn Keels: We did introduce as a part of that same announcement of HP POD 240a – again, the EcoPOD is the nickname – converged systems and cloud systems as well as cloud agile. I'll briefly cover those and then I'll tie it back to EcoPOD. First, we're very proud to say that at its core, HP ProLiant servers help power the world's largest cloud infrastructures with a client roster that includes the three – not two of the three, but the three – most popular social media properties in the U.S., four out of five of the world's largest search engines, and eight out of ten of the world's most trafficked websites.

But beyond these – we call these customers hyperscale or independent cloud providers – we're aggressively introducing private cloud and hybrid delivery solutions in our cloud system portfolio, and that includes burst capacity capabilities between enterprises with private clouds as well as service providers. In fact, we just announced our HP CloudAgile Program designed to streamline not only the technology integration but also the go-to-market motions for these service providers so we can enable a hybrid delivery from traditional IT, to private cloud, to independent cloud, and cloud services provided by service providers worldwide.

At the same time though, a core fundamental trend in cloud deployment is the integration of facilities and IT planning, design, and deployment. That's where the EcoPOD – as sort of the flagship of our POD family as well as the rest of our modular data center offerings – provides that rapid, on-demand, fully configured IT and facilities to be more responsive to the business and capacity needs when service providers, enterprises or those independent cloud providers are deploying converged infrastructure and scale. It's absolutely critical to avoid both over provisioning and under provisioning, and our modular data center portfolio with EcoPODs in the lead there are going to answer that need. Does that make sense?

It totally makes sense. Glenn, obviously this is a major area of focus for you. What is your prediction for the future of data centers? Where do you see the data center going within the large enterprise?

Glenn Keels: Okay. Now, I need to go find my wife and ask for my crystal ball back. I won't presume to be any wiser and have any more crystal balls than the CIOs we serve, but bottom line we do see this seamless integration of IT and facilities planning, deployment and operations as a fundamental shift moving forward, and that's the ongoing benefit of adopting this architecture. Both can be converged infrastructure architecture as a whole and data center smart grid in modular data center solutions as a subset of that.

I'll give you an example. Let's look at a POD customer who was an early adopter. They're not only receiving the benefits of rapid data center deployment, reduced cost and reduced energy efficiency, but they're now poised to tap into the more powerful savings from having those seams, if you will, ironed out between facilities and IT as their business and IT needs grow and change over time. This is a real customer – a large manufacturing firm in Europe that powers its industrial engineering and design on HP converged infrastructure from servers, to storage, to networking and, in fact, PODs. They're currently on our HP ProLiant G6 platforms, hosted in the thousands in PODs and it’s definitely helping them in their product development performance and being more competitive.

But what happens when it's time to refresh that server technology to the latest and greatest next generation ProLiant servers – whenever that might be? For most, this can take months. Even with our HP Factory Express, there's still a significant amount of on-site planning and operational complexity to refreshing those servers generation-to-generation. For this customer, that's been largely eliminated. They already have a concrete pad prepared next to their existing POD infrastructure. When it's time to refresh from one generation of ProLiant to the next, a truck loaded with all of the latest HP innovations and converged infrastructure – all preconfigured and tested to their specification – is going to arrive. The POD will quickly be put in place on top of that concrete pad. We'll add network, power and water. We quickly migrate workloads from one POD to another, and within days, that same truck will take the old POD away. That's incredible value and incredible agility for IT to be able to be more responsive, adaptable, agile, and look out for their bottom line.

The bottom line is we believe in this modular data center approach. Our latest category, this EcoPOD or HP POD 240a, accelerates that adoption of modular data centers to more industries, more customers and more operational models. It fits within our overall family of modular data centers, whether it’s our containerized custom solutions for specialized environments, our chilled water PODs that have a little bit less density and take advantage of existing chilled water or chiller infrastructures, or – on the other side of the EcoPOD – our flex data center, which still has that brick-and-mortar look and feel. You can still have an office in it, a conference room or even a coffee machine. It might take a little bit longer. It's now up to eight to nine months, but it's still rapidly provisioned, energy- efficient and less costly, and that’s all part of this long history of innovation in data center design.

In fact, if you take HP Critical Facility Services and PODs combined, we have more than 43 million square feet of commissioned data space under our belt. Modular is the future.

Modular is definitely the future, and HP's approach makes a lot of sense, both economically and environmentally. I want to thank you for taking the time to share the vision for what's happening with the EcoPOD and HP. Thank you.

Glenn Keels: Thank you.


  • Ron PowellRon Powell
    Ron is an independent analyst, consultant and editorial expert with extensive knowledge and experience in business intelligence, big data, analytics and data warehousing. Currently president of Powell Interactive Media, which specializes in consulting and podcast services, he is also Executive Producer of The World Transformed Fast Forward series. In 2004, Ron founded the BeyeNETWORK, which was acquired by Tech Target in 2010.  Prior to the founding of the BeyeNETWORK, Ron was cofounder, publisher and editorial director of DM Review (now Information Management). He maintains an expert channel and blog on the BeyeNETWORK and may be contacted by email at rpowell@powellinteractivemedia.com. 

    More articles and Ron's blog can be found in his BeyeNETWORK expert channel.

Recent articles by Ron Powell

 

Comments

Want to post a comment? Login or become a member today!

Be the first to comment!