Consolidation, Modernization, and Virtualization: A Triple-Play for Long-Term Enterprise IT Cost Reduction

Transcript of a BriefingsDirect podcast on how IT departments can provide better services with greater efficiency.

Listen to the podcast. Download the podcast. Find it on iTunes/iPod and Podcast.com. Learn more. Sponsor: Hewlett-Packard.

Dana Gardner: Hi, this is Dana Gardner, principal analyst at Interarbor Solutions, and you’re listening to BriefingsDirect.

Today, we present a sponsored podcast discussion on combining some major efforts in IT administration and deployment, in order to both cut costs in the near term and also to put in place greater efficiencies, agility, enterprise business benefits, and long-term cost benefits.

We’re going to be talking about how consolidation, modernization, and virtualization play self-supporting roles alone and in combination for enterprises looking to improve how they deliver services to their businesses. Yet they also play a role in reducing labor and maintenance cost, and can have much larger benefits -- including producing far better server utilization rates -- that ultimately cut IT costs in total.

Here to help us dig into the relationship between a modern and consolidated approach to IT data centers and total cost, we welcome John Bennett. He’s the worldwide solution manager for Data Center Transformation Solutions at Hewlett-Packard (HP). Welcome to the show, John.

John Bennett: Thank you, very much. It's nice to be with you today.

Gardner: As I mentioned, cost is always an issue with organizations and IT departments are among those, facing a lot of pressure nowadays to justify their expenses, show improvements in cost cutting, and, at the same time, improve productivity. John, I wonder if you could help us understand this. We know well enough the cost pressures and economic environment that we’re in, but what has changed in terms of what can be brought to this problem set from the perspective of technology and process?

Bennett: Cost, itself, is easy and complex to deal with. It’s easy to say, "reduce costs." It’s very difficult to understand what types of costs I can reduce and what kind of savings I get from them.

When we look at reducing cost, one of the keys is to get a handle around what costs you're really looking to address and how you can address them. It turns out that many of the cost dimensions can be addressed through a common and integrated approach, building on recent advances in both technology and management and automation tools, on virtualization, and on the investments that companies like HP have been making in focusing on enhancing the energy efficiency and the manageability of the servers and infrastructure that we provide to customers.

This is why, in my mind, the themes of consolidation, which people have been doing forever; modernization, very consciously making decisions to replace existing infrastructure with newer infrastructure for gains other than performance; and virtualization, which has a lot of promise in terms of driving cost out of the organization can increase aspects like flexibility and agility that you mentioned earlier on. It's the ability to respond to grow quickly, to respond the competitive opportunity or threat very quickly, and the ability for IT to enable the business to be more aggressive, rather than becoming a limiting factor in the roll-out of new products or services.

Gardner: We’re certainly well aware of what’s changed in the macroeconomic climate over the last year or so, but what’s different from two or three years ago, in terms of what we can bring to the table to address these general issues about cost. In particular, how we can modernize, consolidate, and get those energy benefits?

Other issues pop up

Bennett: Besides the macro factors around economics that have come into play, we’ve seen some other issues pop up in the last several years as well. One of them is an increasing focus on green, which means a business perspective on being green as an organization. For many IT organizations, it means really looking to reduce energy consumption and energy-related costs.

We’ve also seen in many organizations, as they move to a bladed infrastructure and move to denser environments, that data center capacity and energy constraint, the amount of energy available to a data center, is also an inhibiting factor. It’s one of the reasons that we really advise customers to take a look at doing consolidation, modernization, and virtualization together.

As I briefly touched on earlier, this has been enhanced by a lot of the improvements in the products themselves. They are now instrumented for increasing manageability and automation. The products are integrated to provide management support not just for availability and for performance, but also for energy. They're instrumented to support the automation of the environment, including the ability to turn off servers that you don’t know or care about. They’re further enhanced by the enhancements in virtualization. A lot of people are doing virtualization.

What we’re doing as a company is focusing on the management and the automation of that environment, because we see virtualization really stressing data center and infrastructure management environments pretty substantively. In many cases, it's impacting governance of the data center.

This is why we look at them together. By combining them and taking an integrated approach to them, you not only don’t raise for yourself the issues that some other people may be experiencing, but you can use them to address a broad set of issues, and realize aspects of a data center transformation by approaching these things in an orderly and planned way.

Gardner: We’ve talked about how energy issues are now coming to be much more prominent, cost being a critical issue. Is there anything different about the load, about the characteristic of what we’re asking data centers to do now, than perhaps 5, 10, or 15 years ago that plays into why I would want to modernize and not just look to cut cost?

Bennett: The increasing density of devices in the data-center environment -- racks and racks of servers, for example -- have both increased the demand for power to run them, but, in many cases, have created issues related to cooling from heat in the environment. That has been a trend that has exposed people to risk factors related to energy that they hadn’t experienced before when they had standalone servers or mainframes in the environment.

With virtualization, we also see increasing density and concentration of devices, because you're really separating the assets -- servers, storage and the networking environment --

What we’re doing as a company is focusing on the management and the automation of that environment, because we see virtualization really stressing data center and infrastructure management environments pretty substantively.

from the implications in the business services they are providing. It becomes a shared environment and your shared environment is just more productive and more flexible if it’s one shared environment instead of 3, 4, 5 or 10 shared environments. That increases the density and it goes back to these other factors that we talked about. That’s clearly one of the more recent trends of the last few years in many data centers.

Gardner: I see. So, it’s where we may have had standalone hardware, software applications, siloed or mainframe. When you virtualize, you’re able to distribute the load and therefore look to have much greater ability to increase your utilization generally rather than just at a hit-or-miss basis.

Bennett: Absolutely. I don’t think I could have said it better myself.

Gardner: Tell us a little bit more about green. If we can increase the utilization rates vis-à-vis, what we’re doing with consolidation and virtualization, we also have to look at what we’re doing in the total consumption for electricity and what that means in terms of carbon footprint. Isn’t that possible that we could be looking at ceilings or even regulations on what we can do there?

Capacity is an issue

Bennett: You run into both aspects. Capacity is clearly an issue that has to be addressed, and increasing regulation and governance is as well. We saw the emergence in Europe in the last few months of the Data Center Code of Conduct emerging as a standard for recommending best practices for data centers.

We see an increasing focus in countries like the UK on regulation around energy. There are predictions that that’s going to accelerate in a number of places around the world. So those become part of the environment that data center managers have to deal with, and they can have severe implications for organizations, if they are not compliant.

Gardner: Those have really gone beyond "nice to have" or a way to reduce cost to a "must have."

Bennett: In many cases, that’s very true. Also, there are organizations that had made decisions to be green, where the senior executives and board of directors have made that decision. It’s a management directive and one you have to comply with, independent of government regulations. So, they're coming at you from all sides.

Gardner: I suppose another aspect of this is when you’ve modernized, consolidated, and virtualized your data centers over time, you're further able to automate. You're reducing the amount of labor and manual processes. This strikes me as something that provides an opportunity to manage change better.

Bennett: Yes. When you move to a shared infrastructure environment, the value of that environment is enhanced the more you have standardized that environment. That makes it much easier not only to manage the environment with a smaller numbers of sysadmins, but gives you a much greater opportunity to automate the processes and procedures.

What we see is the infrastructure enabling this. As I mentioned earlier, we're making significant investments in management, business service management, and automation tools to not only integrate infrastructure management with business service management, but also to have an integrated view of physical and virtual resources with line of sight from the infrastructure and the devices all the way up into the business services being provided.

So, you really have full control, insight, and governance over everything taking place in the data center. Many of those are very new capabilities in the HP product suite. Many of these have been announced within the last 12 months.

Gardner: Then, being able to get better automation and standardization across my data center, I should be able to react to the business requirements more quickly.

When you move to a shared infrastructure environment, the value of that environment is enhanced the more you have standardized that environment.

You scale up, scale down, or even shift course better than we would have done in the past.

Bennett: Yes, we use the marketing phrases "flexibility and agility" for that, but what it means is that I no longer have the infrastructure and the assets tied to specific business services and applications. If I have unexpected growth, I can support it by using resources that are not being used quite as much in the environment. It’s like having a reserve line of troops that you can throw into the fray.

If you have an opportunity and you can deploy servers and assets in the matter of hours instead of a matter of days or months, IT becomes an enabler for the business to be more responsive. You can respond to competitive threats, respond to competitive opportunities, roll out new business services much more quickly, because the processes are much quicker and much more efficient. Now, IT becomes a partner in helping the business take advantage of opportunities, rather than delaying the availability of new products and services.

Gardner: These are very important parts of the energy issue, the cost reduction upfront, the ability to be more fleet and agile, and improving the role and responsibility that IT can provide. You won’t have trouble getting people interested in solving these problems, but we get to the point of how we get into a solution, where we can bring these new technological innovations to bear. How do you get started? Where do you focus?

Experience is important

Bennett: How you get started and where you focus really depends on an individual customer and their organization, their capabilities, their staff capabilities, their staff resources, and their experience. Many people are well experienced at doing consolidation projects, and they've been doing virtualization. They have a staff very experienced in looking at things from a business service perspective. For many of them, modernization of the infrastructure, on top of what they've already been doing more aggressively than they’ve done in the past, may be a step to take.

There are certainly tools and capabilities like the Discovery and Dependency Mapping software to help keep an eye on assets and asset configurations, but we are seeing value in being more aggressive in modernizing infrastructure. Typically, people replace servers, for example, on a four to five year cycle, some as aggressively as three but typically four to five years.

In some of the generations of servers that we’ve released, we see 15 to 25 percent improvements from a cost perspective and an energy consumption perspective, just based on modernizing the infrastructure. So, there are cost savings that can be had by replacing older devices with newer ones.

People who have been growing through acquisitions or mergers or for whom individual lines of business control assets on their own, may need to be a little more methodical in building up the picture of just what they have and whether or not they have any -- what some in the industry refer to as -- ghost or zombie servers.

Ken Brill of the Uptime Institute, for example, figures that most people have about 15 percent of their servers not doing anything. The question is how you find out what they are.

Our recommendations to many customers would be, first of all, if you identify assets that aren’t being used at all, just get rid of them. The cost savings are immediate.

If you're going to do consolidation, how do you find out which things are connected to which? For people in that kind of situation, the Discovery and Dependency Mapping software is a wonderful way to go. That’s available for purchase, of course, or it can be delivered from HP services.

They identify all of the assets in the environment, the applications, software they're running, and the interdependencies between them. In effect, you build up a map of the infrastructure and know what everything is doing. You can very quickly see if there are servers, for example, not doing anything.

Gardner: I suppose from that perspective, you can say, We're going to take a couple of spot projects where we know we are going to get a big hit in terms of our return and savings," or "Because of our medium-level solution approach, we're going to start taking out full application sets or sets of services, based on some line of business or geographic definition." Or, we might even go whole hog, if that’s what we're looking at -- more of a data-center modernization to the next generation. All of those seem possible.

Bennett: Our recommendations to many customers would be, first of all, if you identify assets that aren’t being used at all, just get rid of them. The cost savings are immediate. You reduce software license cost, maintenance cost, energy consumption, etc. After that, there are several approaches you can take. You can do a peer consolidation.

If I've got 10 servers doing this particular application and I can have that support the environment by using 3 of those servers, get rid of 7. I can modernize the environment, so that if I had 10 servers doing this work before, and the consolidation gives me the opportunity to go to only to 6 or 7, if I modernize, I might be able to reduce it to 2 or 3.

On top of that, I can explore virtualization. Typically, in environments not using virtualization, server utilization rates, especially for industry standard servers, are under 10 percent. That can be driven up to 70 or 80 percent or even higher by virtualizing the workloads. Now, you can go from 10 to 3 to perhaps just 1 server doing the work. Ten to 3 to 1 is an example. In many environments, you may have hundreds of servers supporting web-based applications or email. The number of servers that can be reduced out from that can be pretty phenomenal.

Gardner: And, all the while, we're reducing physical footprint or the amount of labor required, and we're cutting the energy footprint.

Laying the groundwork

Bennett: All of the above -- and also laying the groundwork for a next-generation data center. We call it an adaptive infrastructure, but the idea is to have this shared resource environment that is virtualized, automated, and capable of shifting assets and putting assets where they’re needed, when they’re needed, and pretty dynamically being able to support growth, pretty seamlessly.

If you take an integrated approach to this by looking at consolidation, modernization, and virtualization together, you actually lay the foundation for that adaptive infrastructure. That’s a real long-term benefit that can come on top of all of the short- and near-term benefits that come with cost reductions and energy savings.

Gardner: We’ve certainly heard about the energy, utilization, and moving to virtualization. They’ve reduced the number of actual servers and therefore the number of people. Do you have examples of some organizations that have gone after these benefits and what sort of experience that they have?

Bennett: We have a lot of examples with people looking to save money. What’s more interesting is to look at a couple of examples of people who have had other objectives, and how they realized those objectives through consolidation, modernization, and virtualization.

An example is a company called MICROS-Fidelio. They provide integrated IT solutions for the hotel industry. They also were looking to improve their competitive advantage and they very specifically were looking at accommodating business growth, even though they had severe limitations in terms of data center space and power capacity. They really didn't want to be investing money in either of those two areas.

They standardized and virtualized their environment using HP blade systems and HP Insight Dynamics. They saw, in terms of business benefits,

If you take an integrated approach to this by looking at consolidation, modernization, and virtualization together, you actually lay the foundation for that adaptive infrastructure.

a 45 percent reduction in missed service-level agreement (SLA) objectives, which meant it reduced the penalties they were paying to their customers by being more predictive in providing better quality of service.

Gardner: In fact, immediate payback.

Bennett: Immediate payback, and not just in terms of cost savings, but in terms of brand reputation. They also had a 50 percent annual growth rate in the data center, which was supported with just a 25 percent increase in IT staff.

They didn’t provide us an absolute dollar figure, but they saved “six figures a year” in personnel cost. This was avoided by being able to do rolling updates to the environment, instead of static updates. Then, they had a threefold faster time in deploying new servers. Again, it was a pretty comprehensive set of benefits, not in just cost savings, but in terms of agility and flexibility, energy, and dealing with space and energy constraints by taking a systematic and integrated approach to consolidation, modernization and virtualization.

Gardner: Before we wrap up, John, I’m really fascinated by this notion of additional automation that the more modern the systems are, the more virtualized, the more ability you have to bring in management capabilities that allow that automation to the almost take off on a hockey stick kind of curve. Not that we want to take people out of the equation, but we want those people to be well utilized themselves. So, what does the future have in store for us in terms of moving the needle yet even further?

Tight control

Bennett: You’ll see improvements in a number of areas. Clearly, at the infrastructure level, we continue to make sure we’re doing everything possible to ensure that the assets themselves are instrumented to be controlled as tightly or as loosely as an organization would like to.

We’re making a lot of investments in ensuring that the physical and virtual assets are managed in a consistent and integrated way, because from a business service’s perspective, the business service doesn’t care where it’s running. But, if you have issues in terms of quality of service, you need to make sure you can track it down through the environment and for that, an integrated view of both is necessary.

Then third, we see the increasing focus on automating standard procedures in business processes and business service management and automation. That has to stretch from the business service down to the infrastructure management, down into the virtual resources, and down into the physical resources. So, it's an ongoing investment in integrating those capabilities, extending the capabilities for the software portfolios, and making sure that that control extends down into the depths of the hardware.

We also continue to make ongoing investments in improving the energy efficiency of the servers, the storage, and the networking devices in the data center. Our first patents in this go back 11 or 12 years now and we continue to see with each new generation of blade system, for example, pretty substantive increases or improvements in the energy consumption and energy demands.

Gardner: Well, great. We’ve been discussing how organizations should consider consolidation, modernization and virtualization as a tag team or a combo team. The payoffs, short-, medium-, and long-term by looking through these different approaches are rather substantial. They're both immediate and have those longer-term strategic benefits baked in.

We’ve been discussing this with John Bennett. He is a worldwide solution manager for Data Center Transformation Solutions at Hewlett-Packard. I truly appreciate your insights, John.

Bennett: Well, thank you very much. I encourage all of those listening to this to take a look at what they can do in their own environments. The potential is pretty significant.

Gardner: Well, great. I also want to thank the sponsor of this podcast, Hewlett-Packard, for underwriting its production. This is Dana Gardner, Principal Analyst at Interarbor Solutions. Thanks for listening, and come back next time.

Listen to the podcast. Download the podcast. Find it on iTunes/iPod and Podcast.com. Learn more. Sponsor: Hewlett-Packard.

Transcript of a BriefingsDirect podcast on how IT departments can provide better services with greater efficiency. Copyright Interarbor Solutions, LLC, 2005-2009. All rights reserved.