For more information on virtualization and how it provides a foundation for Private Cloud, plan to attend the HP Cloud Virtual Conference taking place in March. To register for this event, go to:
Asia, Pacific, Japan - March 2
Europe Middle East and Africa - March 3
Americas - March 4
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 improved data center productivity for a
natural progression toward converged infrastructure. Many enterprise data centers have embraced a shared service management model to some degree, but converged infrastructure applies the shared service model more broadly to leverage modular system design and open standards, as well as to advance proven architectural frameworks.The result is a realignment of traditional technology silos into adaptive pools that can be shared by any application, as well as optimized and managed as ongoing services. Under this model, resources are dynamically provisioned efficiently and automatically, gaining more business results productivity. This also helps rebalance IT spending away from a majority of spend on operations and more toward investments, innovations, and business improvements.
We're here to explore the benefits of a converged infrastructure approach and to better understand the challenges of attaining a transformed data center environment. We'll see how converged infrastructure provides a stepping stone to private cloud initiatives. But, as with any convergence, there are a lot of moving parts, including people, skills, processes, services, outsourcing options, and partner ecosystems.
We're here with two executives from Hewlett-Packard (HP) to delve deeply into converged infrastructure and to learn more about how to get started and deal with some of the complexity, as well as to know what to expect as payoff. Please join me in welcoming our guests today. We're here with Doug Oathout, Vice President, Converged Infrastructure at HP Storage, Servers, and Networking. Welcome to the show, Doug.
Doug Oathout: Thank you, Dana. Happy to be here.
Gardner: We're also here with John Bennett, Worldwide Director, Data Center Transformation Solutions at HP. Welcome back to BriefingsDirect, John.
John Bennett: Thank you very much, Dana. Glad to be with you again.
Gardner: Let me start with you, John. We're talking about some pretty big subjects. There's a lot to chew on here. Data-center transformation (DCT), I suppose, is the most general topic to approach and then to delve down more deeply. What do we mean nowadays by DCT? How does HP define it, and how does that relate to some of the business issues that IT folks are grappling with?
Not one-size-fits-all
Bennett: DCT helps customers implement a data center and infrastructure strategy that's
aligned to their goals and objectives. The key here is that it's customer-driven, and it has to be built around the plans and directions of the targeted organization. This is clearly not a one-size-fits-all type of environment.For many organizations, those strategies for infrastructure can include traditional shared infrastructure solutions or servers using virtualization and automation with shared storage environments. Increasingly, we've seen a natural evolution into a tighter integration of the capabilities and assets of the data center in the fabric infrastructure.
HP's Converged Infrastructure represents a pretty significant step forward in terms of benefits and capabilities for customers looking at having infrastructure strategy aligned to their future needs. The neat thing is that converged infrastructure can be the foundation for private cloud architectures.
Whichever combination of these fits a particular customer, there is the practical challenge of how to change from where you are today to having that implemented? That's what DCT is really about, because it helps implement these strategies through an integrated roadmap of data center projects by consolidation, energy efficiency initiatives, and technology initiatives, like virtualization and automation.
Each of these has its own short-term benefits and returns, but collectively the results get compounded over time, delivering the kind of benefits that we traditionally talk to with DCT. This is all in response to what's going on in customer environments.
I often think of many CIOs as being at the heart of a vise, where, on one side, they have the business pressures. They need to support growth. They need to do a faster job of creating acquisitions. They need to spend more on business projects and innovation. They need to exploit technology for business advantage. They need to reduce costs.
On the other side of the vise are the constraints that they have in the environment that get in the
way of them successfully addressing the business needs -- legacy infrastructure and applications and antiquated methods of managing the infrastructure that make it difficult to be responsive to change, or people with the skills that won’t serve modern technology's needs or environments.Facilities and data centers that were designed and built even five years ago might not have the energy and capacity to support current infrastructure environments. Then, of course, there's energy cost.
So, the CIO is at the heart of this vise. I like to think of DCT and converged infrastructure as kind of the yellow brick road and the Emerald City, where converged infrastructure is Emerald City. It's where you want to get to. DCT is the yellow brick road. It's how you get there, and they complement each other quite nicely.
Gardner: Doug, help me understand why this is important now. The way John is describing it, it seems that the same old approach just won’t hold up, that the trajectory of data centers is unsustainable, whether it's through cost, energy, or capacity issues.
It's not clear to me yet why this converged infrastructure is the right thing to do in totality. Are we talking about a rip-and-replace or are we talking about a gradual direction? Help me understand why, if you are going to move in that direction, you should start now.
Cutting innovation
Oathout: There is a major economic situation going on right now, Dana. As you said earlier,
about two-thirds, if not 70 percent, of the operations budget is spent on maintaining the IT and the IT workload within the data center.When you have a recession, like we just experienced, what happens is that 30 percent spent on innovation or new workload placement gets cut immediately to help manage the budget within an organization. Therefore, in the last 18 months, very little innovation and few new projects were taken on by IT to support new business growth.
Converged infrastructure is important now because we have customers who are starting to spend again and who are starting to see the light at the end of the tunnel. They want their IT environment to be more flexible in the future. So, they're looking at their server and storage upgrades, and how they can implement converged infrastructure, so that the new infrastructure is more flexible and can adapt more to the requirements of the business.
Let me give you an example. A server consolidation using virtualization and new server equipment will generally double or triple your capacity within your data center for the same footprint, just by getting the utilization of the servers up, better performance within the servers, and better capabilities within virtual environments. You can basically double or even triple the size of your capacity within your data center.
As you're going through your technology refresh now, coming out of the recession, you can start implementing better and faster IT equipment.
The same thing holds true for storage. Storage disk drives become twice as dense over a two or three year period. The performance of the drives gets better. So, for the same footprint in your data center you can actually fit twice as much storage.
As you're going through your technology refresh now, coming out of the recession, you can start implementing better and faster IT equipment. You can also use better and more efficient processes -- virtualization, automation, and management. When you put those pools of resources in place, you put them in a virtual environment so they can be shared among applications or can be transferred among applications when needed.
You are in the process now of creating pools of resources, versus dedicated silo resources, like you had prior to the recession, which couldn’t be reused for some of the application, and therefore you couldn’t support business growth.
The opportunity now is to break down those silos, give our customers the ability to share resources in the same footprint they have today, and actually become more efficient, so that when business changes or business needs change, they can adapt to the requirements of the business.
Gardner: So, clearly it's efficiency and better balance between supply and demand of resources, and then being able to apply those resources dynamically with a shared service model. All sounds very well and good. What are the hurdles? What's preventing people from getting to this vision?
Resilient and optimized
Oathout: The big hurdles to get over are the application managers themselves. Line of business comes to the applications team and say they need a SAP deployment or an Oracle deployment, and they tell them what hardware to put it on. In a converged infrastructure environment, you really don’t want to care about the infrastructure you are putting it on. What you want to care about is that it's resilient, it's optimized, and it's modular, so it can grow and shrink with the application's demand.
What you really have is a process change that's required between the IT application managers, the test and development people, and a team that actually runs the infrastructure. They need to talk more about standardization. They need to talk about how their IT comes together.
That's where the Data Center Transformation Workshop that John's team does helps. It gives you an architecture for future deployments, so that you have a converged infrastructure. You have pools of resources to put new applications down or revamp older applications onto a newer architecture, so it becomes more flexible.
You have to break down that silo or break down that fence between application deployments and what line of businesses are telling the application deployers and the people who run the infrastructure. Customers really do see that as a deployment barrier, but they're working through it, because there are significant benefits on the other side, just due to the fact that you increase agility, lower cost, and you have more money and more people to go do the innovation to support the workloads of future businesses.
Gardner: John, it sounds as if we're asking people, in a sense, to rethink things a little bit. Typically, as Doug pointed out, you start with the application set and you deploy it, and then you figure out the best way to operate that over time. We are trying to flip that on its head, thinking about what the operational outcome should be, and then go about applying those applications in the right fashion. Is that fair?
So whether it's 2010 or 2003 or 1992 or mini computers back in 1975, rethinking IT is a very healthy practice, but it always has to be aligned to the question of what the organization and business need.
Bennett: Well, I would suggest that good organizations are always rethinking IT. What are the organization's strategy, goals, and objectives? What is it going to take to realize those objectives? What capabilities do we need from IT in order to make those real? And then, how do we make them happen?
So whether it's 2010 or 2003 or 1992 or mini computers back in 1975, rethinking IT is a very healthy practice, but it always has to be aligned to the question of what the organization and business need.
We also have the question of how it can be exploited for benefit. This is where the partnership between the technology team and the business team comes into play. The technology team will have more insights into how it can be exploited, and the key thing for the business is to make sure they specify their needs and not specify the answer.
Doug characterized it very well, when he said the SAP team wants a new deployment and they tell you what to put it on. The moment you do that, you're losing any of the advantages of a converged infrastructure.
Gardner: As you point out, rethinking IT has been happening for quite some time. We really don't have the luxury of standing still for very long in this industry. On the other side of the equation, you need to have a business or financial rationale to create that change in addition to having the vision of where you would like to go.
So, is there a business case, a rationale, a economic formula of some sort that HP is reflecting about -- how to go to those people who control the purse strings in order to move in this direction of a converged infrastructure?
Business plan in play
Bennett: There clearly is a business plan in play here. A lot of the benefits of this are in the nature of cost savings -- the consolidation, modernization, and virtualization that Doug spoke to -- the savings from energy related projects and investments with Data Center Smart Grid, for example. All are easily quantifiable.
Other benefits have financial benefits too. There's economic return to the organization from being able to roll out a new business service more quickly. There's an economic return to the business from being able to provision more resources when they are needed based on demand, so that demand doesn't disappear. There's a competitive business benefit, which is financial in nature, in being able to respond to competitive threats more quickly.
The business case for transformation and the business case for a converged infrastructure should be constructed, and it's the best way to get the buy in from senior executives.
Technologists playing with toys is not a compelling argument for investment by a business. Technologists making significant investments to make sure that IT is aligned to the needs of the business and having the business case for it is a great way to get approval to go ahead.
Gardner: Doug, when we think about a shared services model and a natural progression toward more of a converged infrastructure that borrows from that shared services mentality, how do we move into this as a manageable progression? How do we avoid that thinking about a rip and replace or a massive disruption or throwing of a switch? How is this applied in terms of a managed process or a progression or evolution?
It also has a very quick payback, because basically you're getting back 30 percent of your disk, which was over-provisioned, and now can be used.
Oathout: The way you go back and look at your infrastructure is very important, Dana. Let's just take the storage environment for a moment. You have a number of storage environments. They could either be direct-attached storage (DAS), network-attached storage (NAS), or storage area networks (SANs).
All these environments have performance application service-level agreements (SLAs), associated with them. If you look at the different types of storage environments, there are different technologies that virtualize those today. These allow you to take large blocks of the storage and put them behind a virtual SAN or behind a virtualized environment, which allows you to share those resources amongst multiple server environments.
For example, we have a SAN Virtual Services Platform from HP, in which you can take heterogeneous storage, put it behind this virtual SAN technology, and actually get 30 percent of the capacity back, because all the over-provisioned disks now become available for all those applications sitting on the other side of the virtual SAN. We have a very similar technology from our LeftHand team, with our P4000, it does the same thing for direct-attached storage.
Using technologies like those to grab the excess capacity you have today by doing storage virtualization is very easy to do. It also has a very quick payback, because basically you're getting back 30 percent of your disk, which was over-provisioned, and now can be used. A lot of our customers don't have to buy disks for two to three months, and then when they do buy disk, they can actually put it behind the SAN environment and multiple applications can use it and share it.
For more information on HP's Virtual Services, please go to: www.hp.com/go/virtualization and www.hp.com/go/services
Server consolidation
On the server side of things, server consolidation is very prevalent today, because the new servers are faster than the old. They have more memory capacity. They have virtual I/O built into them. So, it's very simple for you to consolidate servers, and when you are consolidating you use virtual I/O environment or FlexFabric environment. You then have the capability to dial up and dial down the I/O capacity of the server to meet the demand of the virtual machines running on it.
There is the server consolidation with virtualization that everybody knows, but then there is also the big benefit of storage virtualization and the fabric virtualization that can go on. Those are the three pieces. Once you get them in place, you can then start doing the automation, management, and the provisioning of workloads that John talked about much faster.
It's basically virtualizing that whole environment with resiliency and everything built into our ProLiant boxes and high availability business critical system boxes. You get all the capabilities and all the resiliency you need in them, and then you put virtualization on top of the storage networking and servers, and you really get the pool of resources that you can dynamically allocate.
Those three projects are the ones that give you that base from which you can then springboard your projects or your new applications.
Gardner: We've heard so much in the last year or two about cloud computing and private clouds. I think there is some confusion about private cloud. What we are talking about in terms of these converged infrastructures, the virtualization of various major aspects of your infrastructure, and then getting them to work in some harmony as a fabric. Are we talking about the same thing? Is cloud computing and converged infrastructure essentially the same? What is the relationship?
Converged infrastructure can be for public cloud, private cloud, or for a web workload or an high-performance computing (HPC) workload or an SAP workload. It doesn't really matter.
Oathout: A cloud-computing environment is really an application-rich environment that allows you to bring more users on quickly and expand your capabilities and shrink your capabilities as you need them.
Converged infrastructure can be for public cloud, private cloud, or for a web workload or an high-performance computing (HPC) workload or an SAP workload. It doesn't really matter. A converged infrastructure is the optimal deployment of IT to support any kind of application, because it's modular in nature.
It has the flexibility to have more storage, more memory, less CPUs or more CPUs, less storage, or less memory, but it's all modular, so you can put the pieces together as you need them. So, it is a base support for either a cloud environment or a traditional IT environment. It really doesn't matter. It's designed to support both.
A private cloud is the IT department saying, "I'm now going to create a service catalog for my lines of business to develop upfront." You're getting software as a service (SaaS) now sitting on top of either a converged infrastructure or legacy infrastructure. A converged infrastructure is a lot easy to put SaaS on. But, you make that service catalog available to line of businesses, so they can turn on applications as they need them, very quickly.
Optimizing over time
Then, you can put more users on an enterprise resource planning (ERP) application, an online application, or a Web 2.0 application. IT is there as a support service now, setting that up, taking it down, and optimizing it over time, depending on the business needs.
So, private cloud is kind of that SaaS that sits on either a converged infrastructure or a legacy infrastructure or uniquely designed infrastructures that you get from some of the public cloud providers. Converged infrastructure is the optimal way to develop and deploy that in a standard data-center environment, and it's in support of a private cloud.
Gardner: John Bennett, when we think about that earlier imperative around flipping the balance of spending on operations into spending on innovation, when we think about moving toward a private cloud or cloud environment, and we charge people based on their usage, do they factor together? I'm trying to understand how we can both reconcile moving toward cloud and fabric and "blank as a service," and, at the same time, reduce those costs, so that we can get that business benefit and innovation engine roaring?
Bennett: That's a very interesting question. For an organization to make good business decisions, they need to have a very good understanding, not only of the benefits, which I talked about earlier, but of the costs. In this environment you get line of sight into the cost infrastructure, so you know what it costs you to provide services.
The businesses, in turn, know what it costs to take an offering to market, a cost based on reality and not based on just spread out mayonnaise models of financing. It lets them really understand the business and whether or not it's an investment they should make. There are clearly benefits on that side, if you can go that far. The benefit of moving to that services orientation is that it gives you clear insight into the cost structures.
It allowed a smaller number of people to manage the environment, so that the rest of their IT team could work on improving service levels for the store and how to improve getting new applications to the new environment.
Gardner: I'm always an advocate of showing rather than telling. I hope we have some examples to illustrate how some of your clients have undertaken a converged infrastructure initiative, and what some of the outcomes were. Does either of you have any examples today?
Oathout: There is a retailer we worked with called Stein Mart. They had an inflexible infrastructure to run nearly 300 stores in the Americas, and they were struggling to bring new applications on line quickly enough to support the demands of the store environment.
They bought into the converged infrastructure story. They bought into our BladeSystem Matrix product, which is the combination of storage, server, flexible network, software, and services.
We enabled them to run this BladeSystem Matrix environment. It allowed them to spin up applications in hours instead of days. It allowed a smaller number of people to manage the environment, so that the rest of their IT team could work on improving service levels for the store and how to improve getting new applications to the new environment.
Increased productivity
Stein Mart saw a significant cost reduction, because of the floor space they had in their data center. They saw a significant increase in productivity from their staff. They saw a 2x increase in response time for calls from the stores, and they saw a significant increase in the time to market for new applications. Instead of days, they were taking hours to set up new applications.
The second customer is the Dallas Cowboys. They built a new football stadium in the Dallas area. It's a $1.4 billion investment. In the bottom of the thing is their data center. They run 30 different businesses out of the data center in the Dallas Cowboys stadium.
They have built it on a virtual environment. They have BladeSystems. They have the FlexFabric built into the environment. They went from over 500 servers down to 16 blades, with virtual machines running on them for the point of sale environment within the stadium. It drove a smaller footprint, but also the dynamics in the server and storage environment, so they can bring on new applications for the 30 businesses very quickly.
They changed their infrastructure to support their environment. That's an evolution, versus a Stein Mart, which did a rip and replace to get better productivity to support their business.
Gardner: Any other examples and perhaps ways to demonstrate what HP can bring to this very complex equation?
The infrastructure was there to set up their operating environment on, so that they could run their business relatively quickly.
Oathout: One other example we have is the airport in Dubai, which was a new business, one of the fastest growing airports in the world. They wanted to set up a shared-service environment for their retailers and other businesses around the airport. So, they actually set up a BladeSystem Matrix environment to run their video surveillance, their infrastructure, baggage handling, and all that.
They set up another environment, which allowed their retailers, passport personnel, and other businesses on site to use their shared service environment to really a full service to their client base inside the airport.
So, when a new business, a new government, or a new agency had to come into the airport, they didn’t have to worry about bringing infrastructure with them. The infrastructure was there to set up their operating environment on, so that they could run their business relatively quickly.
Very productive
All three examples: Stein Mart, Cowboys, and Dubai airport, are very productive in how they bring applications online, very reactive to the lines of businesses they are supporting. That's what a converged infrastructure really delivers, besides the lower economic cost that John and I have talked about. It's that efficiency to bring new opportunities to the lines of businesses, accelerate business growth, or increase customer satisfaction.
Gardner: I recall that HP announced converged infrastructure in November 2009, and this is something that I think pulls together a lot of aspects of what HP had been doing for some time. It's a complex process involving people, skills, and different product sets, different professional services, capabilities, and so forth. What makes HP different in terms of how they are accomplishing this notion of DCT, John?
Bennett: What makes us different is that, first of all, we don’t believe one size fits all. We believe that we need to do a good job working with our customers in understanding their strategies and goals and developing an infrastructure strategy that is aligned to that.
We also don’t believe that these infrastructure strategies for the future should have at their core monolithic computing solutions from the past. We also have a very flexible approach in our projects, in that we try to wrap the services we have available around the capabilities of the customer, rather than making them pay to have HP do everything.
Customers who have a great deal of staff, skills, and capabilities with tools . . . will be quite capable of undertaking these efforts on their own.
Customers who have a great deal of staff, skills, and capabilities with tools -- like the Converged Infrastructure Maturity Model and the assessment that goes with that -- will be quite capable of undertaking these efforts on their own.
We try to offer a great deal of flexibility in how we work with customers, and also in how these are implemented. Customers can do these in a traditional customer-owned data-center environment, in an HP hosted environment, or even outsource it to HP. So, there's incredible flexibility based around the customers' needs and interests.
Gardner: You mentioned the maturity model. Is that a potential stepping stone of how to get started on some of these initiatives? Where could some of the folks who are contemplating their next moves architecturally, in terms of transforming their data centers, go to start? How do they get more information?
Oathout: There are two ways to get started. They can contact one of HP’s business partners. Our business partners are enabled to do our Converged Infrastructure Maturity Model. Or, you can come to HP.com/go/ci, and it will take you to the landing page for a converged infrastructure. A client or a customer could click on the Maturity Model, find out about what it can do for them, and then find a practitioner from HP that can come help them, through the Maturity Model, to show them the roadmap or the yellow brick road that John talked about to help them get to converged infrastructure.
Bennett: If the customer is interested in understanding all the things that HP might be able to do with them, they can engage with HP in a Data Center Transformation Experience Workshop, Doug mentioned the CI landing page. They can also go to www.hp.com/go/dct to find out more about that. That will help them take a broader look at the IT infrastructure and facilities and environment, and look at it from a transformational perspective.
Gardner: Focusing on the future, as we look to close up, this strikes me as something that's not just a flash in the pan or a one- or two-year trend. This seems to me a long-term trajectory. This is pretty much an inevitable way in which data centers are going to develop, that is to say converged, fabric, service-oriented, with the efficiency of dynamic provisioning involved. Any thoughts about where this direction is going to take us, and do you agree that this is essentially inevitable?
Economies of scale
Oathout: It is inevitable, just because of the economies of scale, Dana. Truly, when you start bringing a storage and server and networking platforms together through a flexible fabric, the economies of scale of a shared resources and open systems is going to drive down the cost of acquiring IT. Then, with the software and the services capabilities that companies bring to market, they're going to bring the efficiencies along with them.
So, it is inevitable, starting with the simplest of workloads, moving to some of the hardest of workloads, that you are going to have a converged infrastructure. You are going to have application as a service, whether it's internal or external from a cloud provider, just because the economies of scale are there, and the ability to deploy the stuff is so simple once you get it set up that the efficiencies are also there besides the economies of purchase.
Gardner: Any last thoughts John in terms of the future direction and how long of a trend we are talking about here?
Bennett: How long a term is always difficult to say. One of the exciting things about IT in general is that we see this wonderful yin and yang, this give and take between technology advancements and customer expectations and uses. Customers challenge us to step forward to meet tomorrow's problems. Technology evolves, and we challenge customers to take advantage of it for a business benefit.
That's going to continue, as Doug highlighted, the economic value that comes from this convergence of infrastructure, and the economies of scale are very compelling, but I'm not going to predict how long it's going to last.
Gardner: Well, we'll certainly find out, won't we? It's been very good speaking with you. We've been talking about improved data center productivity through a progression to converged infrastructure.
We've been joined by two executives from HP. Doug Oathout is the Vice President of Converged Infrastructure at HP Storage, Servers, and Networking. Thanks so much, Doug.
Oathout: Thank you Dana.
Gardner: Also John Bennett, Worldwide Director, Data Center Transformation Solutions at HP. Thanks again, John.
Bennett: And thank you, Dana, and thank you, Doug.
Gardner: And, thank you all for listening. This is Dana Gardner, principal analyst at Interarbor Solutions. You've been listening to a sponsored BriefingsDirect podcast. Thanks for listening and come back next time.
Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. Learn more. Sponsor: Hewlett-Packard.
For more information on virtualization and how it provides a foundation for Private Cloud, plan to attend the HP Cloud Virtual Conference taking place in March. To register for this event, go to:
Asia, Pacific, Japan - March 2
Europe Middle East and Africa - March 3
Americas - March 4
You may also be interested in:
No comments:
Post a Comment