VSM interviews Stephen Pollack, CEO and President of PlateSpin By VSM News Staff published: Thursday, June 24 2004
Recently PlateSpin Ltd., a leading developer of solutions that bring flexibility and automation to the multi-architecture data center, announced the availability of PlateSpin PowerP2V™, the next generation platform for converting servers between physical and virtual machine architectures.
The new PlateSpin PowerP2V™ platform allows data centers to dramatically accelerate their physical to virtual server consolidation programs by eliminating manual tasks, and reducing human error through automation.
Stephen leads the overall business strategy and product management activities for the company. Stephen is a seasoned IT veteran, with a strong history of creating products and services that exceed customer expectations. He brings over 20 years of experience in marketing, sales, development and lifecycle support of successful commercial software products in a variety of technology and market segments.
Good morning, Stephen. Before we talk about PlateSpin PowerP2V™, can you tell us more about PlateSpin and how it’s positioned in the server virtualization industry?
As a company we were founded in March 2003. Our focus in the market has been on delivering flexibility and automation solutions to the multi-architecture data center. By that we mean organizations that have adopted different architectural approaches to how they host certain IT business services. Some are based on older computing models, pre-WinTel days, some based on combinations of Windows and Linux, and more recently, new resource architectures such as virtualization, which people are starting to think about now.
So one of the essential themes to our company, and the way we’re moving forward with our products, is to make sure we can cater to the needs of a multi-architecture data center. In today’s terminology it largely focuses around server virtualization from companies like VMware and soon Microsoft with Virtual Server, but the general goal is to provide the capabilities that we have across a variety of architectures, beyond even just Intel platforms as well, in order to provide a single point of management and automation capability for enterprise data centers.
Tell us about your most recent announcement, PlateSpin PowerP2V.
We initially added a physical-to-virtual or P2V capability to our first product, PlateSpin Operations Management Center (POMC), in June of 2003. At that time our main goal was to capitalize on the fact that we had the ability to see into the physical side of the network, not just the virtual side, and therefore could create a highly automated way of moving server configurations across that boundary.
So we extended the POMC product with a P2V module. But it carried with it some of the overhead and some of the architecture that went along with the full management product. Which really meant overhead around the way it was set up, overhead around the use of agents, and was in some sense intrusive into the network especially if your predominant use of the product was the P2V module.
So we saw a need to separate that functionality into its own product offering, and start to provide conversion capability in a less disruptive fashion to the data center catering to two principal needs. One is service providers, like Expert Server Group, who go into accounts and help companies adopt new data center infrastructures based on virtualization, and need a solution that they can come in with, and in a very touch-less manner help people convert their physical environments to virtual machines without any disruptions. Unlike other approaches, there is no network re-configuration, no use of software agents, really no need to gather any information from the user or otherwise harm the environment. With PowerP2V you only need the right authority or server login credentials to do the conversion.
PowerP2V is the first product based on a new architecture that we started implementing when we formed the company, to bring together what we think is the next generation of capabilities for not just our product, but for the needs of the data center as we saw them unfolding.
The initial focus of PowerP2V is to help people implement projects like server consolidation, where you go in either as a service provider or the IT person in the data shop, and by running the tool on your laptop or desktop you can reach into the network, grab a physical server, and move it over to a virtual machine without visiting the source or target servers.
As we see the need unfolding, we understand the requirement to have general value to the enterprise on an ongoing basis. Even though you might ultimately finish all of your physical-to-virtual conversions, you always have a need to move things around and reorganize, and consolidate your data center. So PowerP2V for us is a long-term platform that customers can invest in to gain greater flexibility for the environments they’re managing.
Today it might need to be physical, tomorrow it might be virtual based on VMware, maybe six months down the road it needs to be a different kind of virtualization, maybe it needs to go back to physical. The PowerP2V platform would let you move that environment around, and provide a lot of flexibility around how it’s configured and re-configured as it takes each step through its life cycle.
It sounds like one of the key differentiators for your PowerP2V product vs. some of the other P2V tools is that yours is non-invasive.
Yes, in fact I would say that on first blush that is probably the most significant enhancement and value of the product compared to every other solution. It does not make use of any third party software. You do not need imaging, you do not need temporary disk space, boot CDs, service packs or anything to do the conversions, which is pretty typical of the other products that are out there.
Second, there is no need to change the network, so you don’t need to enable PXE on your servers, you don’t need to reconfigure your VLANs. You just reach into the server, capture its content and automatically manage all the issues on the destination side all the way down to creating the virtual machine.
If you contrast it with some of the earlier solutions people have had, the IT person would normally have had to create the empty virtual machine in order to do the conversion. With PowerP2V we also take care of all of that. So as an IT person you don’t need to know a lot about virtualization in order to do a conversion. You really only need to understand what you want to do with basic OS configurations such as IP settings and hostnames.
That’s definitely a key feature and a key differentiation for us, the non-invasive nature of it.
Does Power P2V support both Windows and Linux?
Yes it does. That’s another first for us. We believe we’re the only vendor today who can provide the same level of automation around converting Linux as we can with Windows. We have the full breadth of Windows coverage and we’re expanding Linux to include Red Hat, SUSE and other forms as those opportunities arise.
But even beyond that, we’re looking to support Netware as another platform. There are a number of companies who have large Netware environments and they’d like to preserve those. But the only viable way to do it is with virtual machines, because increasingly the hardware vendors are not providing support for Netware on the newer platforms.
What stage are you at with PowerP2V now? Is it in general beta, or general availability?
We ran a beta program for about 2 1/2 months, which ended at the beginning of June, and did a final certification pass on the product to deem it generally available. So we are now in a fully released mode having already made a good number of sales, canvassing for new leads and developing customer relationships, and starting to roll it out into the market.
Are there other key differentiation points with PlateSpin’s P2V solution?
I think people underestimate the value of the level of automation that needs to be provided. When you are only converting 2, 3, 4, 5 servers you could do it manually, although that can have a high degree of error and time associated to it. But even if you are using a semi-automated tool you still have a lot of IT activity, a lot of involvement from the IT person.
If you’ve seen a demonstration of PowerP2V, it’s literally 1-2 minutes of IT time, start to finish, to get a conversion going. And unless you have nothing else to do and you want to watch it, you really don’t have to do anything else. You can go off and do other IT activities, or you can even start other conversions, because the product fully supports concurrency around how conversions work. I can take five servers, and start converting them to a single virtualized box all at the same time. Within about 10-15 minutes of IT time I can have those conversions all running. So a high degree of automation we think is very important.
Another differentiation which may be more subtle in the market today is that the product as a platform has an API associated with it. We introduced that for several reasons. One is that a number of people asked us early on about the ability to script the conversion. Instead of using the user interface they’d want to put it in a batch file and attach it to some other work flow or processes that they have. So there is an API available that would only be possible if we could deliver the high degree of automation.
Similarly there are other vendors out there who are in the systems management space and they want to add server conversion as part of their product set. They can use the API to integrate PowerP2V functionality into their product. The notion of an API is now a central part of our overall product offering, all the way through to the management functionality that we’re adding over the next many months.
When you sum it all up, non-invasive issues are important, the high degree of automation, the API, the coverage of Windows and Linux, but also all the flexibility as well. We’ve run into a number of integrator projects where one of the key goals is to really change the environment as I convert it. Not just virtualize, but I need to redo my NT 4 configuration, because I did it with a C drive, a D drive, an E drive, because that’s what the right thing was back then. But that’s really silly, I need to expand all my drives, and I want to do that during the conversion.
So we support all that flexibility as well, so you can effectively right size your servers as you convert. And when you look downstream, it’s all set up to go multi-directional, so you can do physical-to-virtual, virtual-to-virtual, virtual back to physical, physical-to-physical, and it’s therefore an excellent platform for customers to invest in now. It will continue to add value to them as they go forward.
Stephen, can you characterize PlateSpin’s relationship with VMware?
It’s been an interesting relationship. We worked hard to create an added-value relationship between our products and VMware’s own products including VirtualCenter. We are working with VMware to create an ISV program for vendors like ourselves. Their current formal partner programs are more related to their distribution channel, not necessarily people building applications like us that complement the VMware technology.
So we have a very active relationship, but not a formal partnership. Which is something that VMware is looking to add to their partner program over the next little while. So we’ve been working with them on that, and expect some announcement shortly on that level of partnership.
What are the operation management capabilities of PlateSpin’s solutions?
The interesting issue is what’s really important to the data center over the next few years? One of the things that we see some companies following is that they are starting to build up a different management solution and people infrastructure for each part or architecture in their data center. We would argue that trend will proliferate the problem people have managing data centers, where they have different solutions that don’t interoperate. So one of the key features of the product family is that what we have is from the ground up designed to manage, support, discover, provision across the virtual and physical environments.
Unlike any of the other products out there, we believe currently we are unique in this regard. PlateSpin’s Operations Management Center can manage physical servers, maybe one runs the domain controller and one runs the database and can manage virtual servers that each have 4, 5 or 10 or however many VMs on them, and you can manage them as a single entity with a common paradigm. We don’t see the virtual servers as physical servers like other products do. We manage them as virtual machines natively to make use of their specific advantages.
You can do software distribution across the mix, patches, updates, new releases, health management, basic provisioning to either create new physical servers or new virtual machines, all with the same flexible provisioning paradigm. So for one IT person who is just now engaging in virtualization, they don’t have to adopt a different solution set just to manage that. What you really need is a solution that has within it the underpinnings of how to manage a virtual environment that doesn’t really need a different paradigm for it. We brought the two together in our product, and with the single solution you can manage Linux, Windows (and Solaris) together, which is very important to the data center.
We will keep evolving that product to add additional automation, to strengthen what we call business service management, which is more the way of mapping the servers that are in the environment back to the business purpose they provide, either for billing purposes or for organization and analysis.
Where does PlateSpin think virtualization technology is going in the next few years? I’m sure you, like many companies, see that we are only at the beginning of this revolution. Can you share your vision?
If we look at the trends right now, it’s very interesting from my perspective. Like many people I talk to, you read a lot about the future of computing. Many of them hinge on a common theme, which is that the concept of the computer has really been pushed into the fabric of the society. We see that today with the Internet. But computing itself as a technology hasn’t yet been pushed down like telephone and electricity have.
So in many ways this seems to be the tip of the iceberg, where computing is now going to fundamentally change. And the notion of the computing resources that are needed to run a business application are going to start pushing themselves deeper and deeper into the fabric of the way a basic business runs. Products like VMware, Microsoft Virtual Server, are early ways to go down that path in a very non-disruptive fashion compared to, say, adopting blade computing, which might be slightly more disruptive in some companies.
If you look out further, there are clearly enough technology pieces in place now, or being released shortly, that you literally can fully separate from an application the resources it needs to run: memory, disk, CPU and network. And if you take that to its logical conclusion, it might be reasonable to assume that in a typical business there will be a big box somewhere that has everybody’s disk, CPU, memory, network resources, and you’ll run applications remotely and they’ll just draw on that as they need.
That would be many, many years out, of course, if it happens quite that way. But that’s an interesting trend for us, because one of the challenges people typically have is how we adopt to that infrastructure. How do we go down that path? And one of the areas that we are heavily focused on is helping companies go down that path by trying to tackle the problem from two sides. One side is this whole notion of infrastructure conversion. How do I help you bring in these new technologies? How do I help you convert your existing data center into those new architectures? And then once you’re there, how do I keep them below the fabric, so that I can manage them consistently across the different architectures? And then when I get to that end point I’ll be where I want to be.
So whether you look at the approaches IBM is taking, or HP, or Microsoft, they’re largely all trying to accomplish the same thing. To create those degrees of separation between the business applications and the things that they need to run successfully, and build up an infrastructure – hardware, operating systems, applications, communications standards like DCML or Microsoft DSI, IBM’s WebSphere model – all tie together to realize this future goal.
PlateSpin is working both sides of the equation, and to some degree feel that there is not enough emphasis right now being placed on how to do migrations. So we’re really enthused about how well products like PowerP2V have been received as a first step in helping companies go down that path.
What is the best way for companies to find out more about PlateSpin?
There are two ways. Directly through our Web site, www.platespin.com, or, as readers may know, we announced the next level of expansion in our business partner program. We do most of our work through third parties who are also engaged in helping companies solve the same problems that we are, and carry our products for resale or use them to deliver service. If people are interested in our products, they can either contact a partner directly, or contact us, and we’ll help organize a relationship between them and one of our partners. That’s really the best way. We have some materials on our Web site as well that people can download. There is a white paper and product data sheets. We can also arrange live demos of both products, which is an effective way to make a purchasing decision.
|
