Transcript of a BriefingsDirect podcast on the improved business metrics from adopting a virtualized storage architecture.
Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. 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 storage virtualization. You've heard a lot about server virtualization over the past few years, and many enterprises have adopted server virtualization to improve their ability to manage runtime workloads and high utilization rates to cut total cost.
But, as a sibling to server virtualization, storage virtualization has some strong benefits of its own, not the least of which is the ability to better support server virtualization and make it more successful.
We're here to discuss how storage virtualization works, where it fits in, and why it makes a lot of sense. The cost savings metrics alone caught me by surprise, making me question why we haven't been talking about storage and server virtualization efforts in the same breath over these past several years.
To help us explain how to better take advantage of storage virtualization, we're joined by Mike Koponen, HP's StorageWorks Worldwide Solutions marketing manager. Hello, Mike.
Mike Koponen: Hello, Dana. How are you doing today?
Gardner: Doing very well. Thanks for joining us.
Koponen: You bet.
Gardner: As I said, a lot of folks have been taking up more server virtualization and understanding its benefits. It's become quite popular, particularly in the down economy, where cost is so important. Storage virtualization offers a number of the same types of benefits. Tell us why storage virtualization makes so much sense.
Koponen: Dana, you mentioned that, particularly in today's economic environment, customers need to boost efficiencies from their existing assets as well as the future assets they're going to acquire and then to look for ways to cut capital and operating expenditures. That's really where storage virtualization fits in.
It's a way to increase asset utilization. It's a way to save on administrative cost, and it's also a way to improve operational efficiencies, as businesses deal with the increasing storage requirements of their businesses. In fact, if businesses don't reevaluate their storage infrastructures at the same time as they're reevaluating their server infrastructures, they really won't realize the full potential of a server virtualization.
Gardner: A few years ago, people were putting in servers as fast as they could. Basically, their goal or their motivation was simply to keep up with demand. I have to believe that's the case with storage as well, and storage requirements are still growing rapidly. How do you both keep up with the high demand for more and try to cut cost at the same time.
Koponen: It's an excellent question and one that businesses deal with all the time. As you say, the storage requirements aren’t letting up from regulatory requirements, expansion, 24x7 business environments, and the explosion of multimedia. Storage growth is certainly not stopping due to a slowed down economy.
Storage virtualization and server virtualization are tools that businesses are using to deal with those. In the past, as you said, customers would just continue to deploy servers with direct-attached storage (DAS). All of a sudden, they ended up with silos or islands of storage that were more complex to manage and didn't have the agility that you would need to shift storage resources around from application to application.
Then, people moved into deploying network storage or shared storage, storage area networks (SANs) or network-attached storage (NAS) systems and realized a gain in efficiency from that. But, the same can happen. You can end up with islands of SAN systems or NAS systems. Then, to bump things up to the next level of asset utilization, network storage virtualization comes into play.
You can pool all those heterogeneous systems under one common management environment to make it easy to manage and provision these islands of storage that you wound up with.
Gardner: You mentioned this notion of silos of storage and I think I heard at least two or three different levels of silos of storage. Can you break that out for us? What are we really talking about, when we think about the various components at play here?
Koponen: I break it down into three levels. One, I'd call basic virtualization. That's where you just have internal storage in your servers or direct attached storage to those servers. The next level would be what I'd call virtualized network storage. We've got SAN systems that have the ability to virtualize the arrays and the disk spindles within that SAN system.
The third level is what I call network-based storage virtualization. There, you have the ability for heterogeneous storage systems to all be managed under a common structure and virtualized as a single common pool of storage. Those would be the three levels that I break them down into.
Gardner: So, the goal with storage virtualization is not just to virtualize on each of those levels, but to virtualize them all together, so there is a single pool of storage. Is that correct, or that I am oversimplifying?
Koponen: No, that's basically it. In the second two levels I described, where you've got a SAN system, those can also come in two types. You can have a traditional one that's non-virtualized and then you can have a virtualized one, such as the HP Enterprise Virtual Array or the HP LeftHand SAN, where you have the ability to stripe data out across disk spindles and multiple drive trays, and all of that is abstracted from the system administrator.
The storage is virtualized, and then the level above that is where you have network-based storage virtualization, such as our SAN virtualization services platform, that can take heterogeneous storage systems, multiple SAN systems for multiple vendors, and present those as one common pool of storage. It's this concept of pooling storage, but at different levels.
Gardner: Of course, it's a big management task to be able to do that and then get to the storage the way that you want to prioritize different storage requirements and some responses based on the application set or whether you're doing it for backup or archives. Is that right?
Koponen: That's true. There are different needs or requirements that drive the use of storage virtualization and also different benefits. You mentioned some of them. It may be flexible allocation of tiered storage, so you can move data to different tiers of storage based upon its importance and upon how fast you want to access it. You can take less business-critical information that you need to access less frequently and put it on lower cost storage.
The other might be that you just need more efficient snap-shotting, a replication of things, to provide the right degree of data protection to your business. It's a function of understanding what the top business needs are and then finding the right type of storage virtualization that matches those.
Gardner: It also sounds like we're taking a complete look at storage. We're looking at it from all angles and, therefore, are able to architect in such a way that we can take advantage of all the capacity we have and do that intelligently. Is that a fair assumption?
Koponen: That's true. One key driver is boosting asset utilization. We found that in a lot of businesses they may have as little as 20 percent utilization of their storage capacity. By going to storage virtualization, they can have a 300 percent increase in that existing storage asset utilization, depending upon how it's implemented.
Gardner: Mike, tell me how this relates to server virtualization. If I've got a server virtualization program underway and I've enjoyed some benefits from that, what is taking this added step to storage virtualization going to do for me?
Koponen: Well, a couple of things, Dana. First, in order to take advantage of the advanced capabilities of server virtualization, such as being able to do live migration of virtual machines and to put in place high availability infrastructures, advanced server virtualization require some form of shared storage.
So, in some sense, it's a base requirement that you need shared storage. But, what we've experienced is that, when you do server virtualization, it places some unique requirements on your storage infrastructure in terms of high availability and performance loads.
Server virtualization drives the creation of more data from the standpoint of more snapshots, more replicas, and things like that. So, you can quickly consume a lot of storage, if you don't have an efficient storage management scheme in place.
And, there's manageability too. Virtual server environments are extremely flexible. It's much easier to deploy new applications. You need a storage infrastructure that is equally as easy to manage, so that you can provision new storage just as quickly as you can provision new servers.
Gardner: So, is this a case of a whole being greater than the sum of the parts? If we do server virtualization well and then we do storage virtualization well, not only do we get the usual benefits in terms of capacity, cost, flexibility, and intelligence at each of those perspectives, but, by combining them, we get something additional.
Koponen: Yes, you certainly do. The way I would describe that X factor of what you're getting in addition is just the highest level of business agility and flexibility. The underpinning of that would be that you're making maximum use of your assets, both your server assets and your storage assets.
Gardner: Is there something else here in terms of security, compliance, complexity, or those other necessary things to deal with nowadays? Do we get anything else in combining these two?
Koponen: You certainly get an increased degree of data protection by being able to meet backup windows and not having to compromise the amount of information you back up, because you're trying to squeeze more backups through a limited number of physical servers. When you do server virtualization, you're reducing the number of physical servers and running more virtual ones on top of that reduced number.
You might be trying to move same number of backups through a fewer number of physical servers. You also then end up with this higher degree of data protection, because with a virtualized server storage environment you can still achieve the volume of backups you need in a shorter window.
Gardner: So, it's better control, better understanding, higher utilization, and lower cost. If someone is interested after hearing this, where do you start, how do you undertake a journey? I assume you don't do this all at once, but rather it's something you need to do on a rollout basis. Where do you start when it comes to storage virtualization?
Koponen: Step one is assessing your environment and understanding what your starting point is going to be. Is it a greenfield environment, where you've got a lot of departmental, work-group type servers that you don't have tied into shared storage or virtualized storage? It might be starting with putting in place virtualized storage to support those.
Or, do you have existing SAN systems in place that are just underutilized. Then, you might look at putting in place, say, the HP SAN Virtualization Services Platform (SVSP), to get a higher degree of asset utilization out of the existing systems.
It depends on where you're starting from. So, step one is to determine that, figure out where your most underutilized assets are, and what's causing you the most pain today from a management complexity standpoint. Or, it could be the case that you don't have an adequate business continuity plan in place. That's your key factor in where to start. So, it's assessing that starting point, Dana.
Gardner: Let's drill into that business continuity one for a second. That's pretty important. What does virtualizing your storage bring to the table, when it comes to data recovery, disaster recovery, backup, archiving, or continuity issues?
Koponen: Well, first, when you virtualize your servers, you're taking multiple applications and running them on a single physical server. You're more exposed now to that single physical server going down, because, if that single physical server goes down, you've lost multiple applications, and not just one. So, the need for high availability goes up.
Server virtualization suppliers like VMware, Microsoft, and Citrix, all have capabilities to provide high availability on the application side. You need to make sure you match that with high availability on the storage infrastructure side, so that you've got the same capabilities within your storage from a high availability standpoint as you do your sever infrastructure.
Gardner: Right, it doesn't make sense to have the applications humming along at whatever requirements are, if the storage and data can't keep up.
High application availability
Koponen: Exactly. From an HP portfolio standpoint, we have some innovative products like the HP LeftHand SAN system that's based on a clustered storage architecture, where data is striped across the arrays and the cluster. If a single array goes down in the cluster, the volume is still online and available to your virtual server environment, so that high degree of application availability is maintained.
Gardner: Mike, how about some examples? For folks that have done this already, what are the typical scenarios? What are some of the paybacks? What's the usual case scenario?
Koponen: Dana, there was a white paper recently done by IDC on the business value of storage virtualization. It looked at a number of factors -- reduced IT labor, reduced hardware and software cost, reduced infrastructure cost, and user productivity improvements. Virtualized storage had a range of payback anywhere from four to six months, based on the type of virtualized storage that was being deployed.
It found asset utilization increases up to 300 percent, savings of administrative cost of 2x to 3x, and shrinking back-up times by up to 80 percent as well. The benefit in the payback was really compelling. That IDC paper is posted on the HP website.
Gardner: What are some of the business returns? Clearly, we've got some cost benefits and technology benefits that folks in the IT department would enjoy, but what would we expect from storage virtualization for the larger business outcomes or goals?
Koponen: You have these benefits of reduced CapEx and OpEx that companies can take to the bottom line, particularly in these economic times, and you also have improved business agility as well. Let's say, a company makes an acquisition and they've got to merge an existing IT resource into their existing IT infrastructure. The ability to do that is going to be greater, given that you've got a virtualized storage infrastructure in place.
Gardner: That gives more agility for changing your organization from a merger and acquisition perspective. How about sourcing, when it comes to what we think about now as cloud computing? Is there some benefit in having virtualized storage that gives you more options for your sourcing?
Koponen: Well, it gives you more options in terms of the flexibility with which you manage your internal cloud, how you can meet your quality of service levels to your internal user community, and how you partition out storage to them, because you can make much more efficient use of those resources. Using your outsourced cloud providers, you can augment that existing server and storage capacity that you've got in place.
Gardner: Returning to the road map of how you would get started and involved with this, what else do you need to consider other than certain technologies? Is this something that's going to change the nature of your organization? Are we going to be asking different folks inside of IT, and perhaps outside of IT, to work together in ways they hadn't before? What are the cultural implications?
Koponen: That's a good question, Dana. In the case of enterprise organizations, you may have the storage management done by a particular set of folks. Then, you may have the management of it, the server infrastructure, done by another set of folks. This will bring those two sets of resources together and provide them more efficient platform to be able to work together.
In medium-sized businesses, it's all about being able to manage storage assets without having to have expert storage administrators in place, so that server administrators can manage the storage assets as easily as they do their server assets.
Gardner: I wonder, if there's something we left out. Is there another item around this that folks should be aware of?
Koponen: I don't think so. However, for people who want to learn more about storage virtualization and what HP has to offer to improve their business returns, I suggest, they go to www.hp.com/go/storagevirtualization. There they can learn about the different types of storage virtualization technologies available. There are also some assets on that website to help them with the justification of putting storage virtualization within their companies.
Gardner: Just to be clear. When HP approaches storage virtualization, you're working with a number of different vendors and suppliers and different technologies. This is really quite a heterogeneous landscape. Is that correct?
Koponen: That's correct. HP has very strong relationships with all of the server virtualization suppliers in the marketplace, so that we can bring complete solutions to bear on customers. We use best-of-breed technology from the entire ecosystem.
Gardner: Well, thanks. We've been hearing about server virtualization in the past few years, but today, we've taken the time to look at a sibling, storage virtualization. It involves improved runtime workloads in the server side, getting the storage and data that it needs, but there is also a lot of pure, economic rationale for going about the storage virtualization on its own.
Here to help us better understand the whys and hows of storage virtualization is Mike Koponen. He's the HP StorageWorks Worldwide Solutions Marketing Manager. Thanks for your time, Mike.
Koponen: Thank you, Dana.
Gardner: 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.
Transcript of a BriefingsDirect podcast on the improved business metrics from adopting a virtualized storage architecture. Copyright Interarbor Solutions, LLC, 2005-2010. All rights reserved.