PowerP2V: Following up with Stephen Pollack of PlateSpin By VSM News Staff published: Wednesday, November 24 2004
VSM followed up with Stephen Pollack, CEO of PlateSpin, after VMworld and the very successful November 23 Webinar on PowerP2V.
VSM: What’s new at PlateSpin?
SP: The VMware VMworld user conference confirmed what we’ve seen, that virtual infrastructure seems to be arriving as a real mainstream alternative to other computing models. We definitely saw an upswing in our business over the summer, driven by customers moving into more aggressive planning and implementation around how they can save money using virtual infrastructure. One of their key requirements is to accelerate the migration or the divestiture of their physical infrastructure into virtual infrastructure. That’s where our PowerP2V product has taken hold quite strongly in the market as a leading companion solution to VMware and Microsoft Virtual Server.
Our own experience at VMworld, where it was clear that anybody doing work with virtual infrastructure would be attending, was that we seemed to be the popular booth as prospects stopped by to talk and explore what’s available to help with their virtual infrastructure deployments.
We are addressing a very specific need that everybody recognizes they have, no matter who they are or how many servers they have. Unlike a few years ago, when you could bring in new technology and build or rebuild your business applications to use that technology, companies can’t afford this anymore. As such, a lot of companies are challenged by how they migrate their infrastructure forward to new technology. There is a lot more emphasis on how to accommodate what is already in the data center, let alone build new stuff.
VSM: That’s becoming more and more the case, especially as its not just extremely large companies that move into these technologies, but the mid-market too.
SP: Our leads have ranged all the way from people with three servers in an old NT 4 environment who want to use PowerP2V to move them to a new machine, to a couple of our customers who have 5,000 servers or more and a multi-year plan to use virtual infrastructure as much as they can. Their needs are the same as the small company’s needs, just on a much bigger scale. They’re looking for savings through the remote access PowerP2V provides, the high degree of automation it provides, as well as all the configuration flexibility that the product now offers.
We see PowerP2V as a horizontal offering into the market; it is not related to any specific industry segment. As a software company, it bodes well for our future to be leading with a solution that meets a horizontal industry need because there is maximum opportunity within the target for us to explore.
VSM: Are you getting global business?
SP: Yes. As a software company it’s been much easier over the last few years to reach the global market because of the Internet and what it offers for basic communication for sales, support and marketing especially. We’ve also been successful in building a network of Value Added Resellers and systems integrators all across the world: South Africa, Australia, all over Europe, Brazil, Mexico, Spain, UK, US, Canada.
That has forced us to invest in some aspects of our business that we may not have done five or ten years ago, where you wouldn’t typically start global this way. You’d target the local market, then the nearest geography would be the rest of North America, then perhaps in Europe. Now it’s really all happening at the same time. We have as much interest in, and as many customers, across the major territories as we do in any one territory.
VSM: That seems to be a new business model now, especially for software, which is not always a tangible, in-the-box product.
SP: Exactly, you can Web download products like PowerP2V so you don’t need the brick and mortar business structure of other markets. Looking at it from the data center point of view, the PowerP2V product specifically is one that is easier to move through a channel than a big manufacturing system, for example, or a complex data center management system.
VSM: What was your experience at VMworld like?
SP: We understand there were 1,800 people attending. There were about 25 vendors, and we had an extremely busy booth. We had six people at the show, and all six were needed at the booth for the whole show. We were constantly demonstrating our products, and previewing some new technology that we’re planning for next year. We actually sold some software at the booth! It is an indication that some momentum has been established globally in the virtual infrastructure space.
VSM: Can you share some of the new technology?
SP: We focused on previewing three innovations. We had one demo station with the latest release of PowerP2V, Version 3.6, which added some additional capability to make it more useful in a large-scale company network.
We also previewed the industry’s first V2P capability, the ability to de-virtualize your environment, which seems to be a next need for companies that are deploying virtual infrastructure in the testing and development labs but not necessarily everywhere in production. They can use PowerP2V to take the production server, whether it’s a VM or physical, migrate/clone it into the test lab to do the change management exercise, and when they’re done with it they can de-virtualize it or do a virtual-to-virtual copy and put it back into production. We were previewing the first version of that.
We believe we are the first in the world to prove that you can do virtual back to physical migrations. It will be released in early 2005 as an extension of the PowerP2V product.
VSM: That makes sense, especially in Test and Dev. When you’ve got it right you might want to roll it out to production.
SP: Yes. And there’s the background issue of whether or not you can get support for Windows in a VMware VM, so people are looking at V2P as an alternative way to get support for Microsoft.
Another growing interest for virtual infrastructure is Disaster Recovery. We see clients exploring maintaining a disaster recovery environment as a collection of virtual machines. You have a lot less hardware required for sure, but when you need to recover, PowerP2V would offer you both alternatives, either activate the VMs and run things that way, or procure bigger hardware and run the disaster recovery alternative on a bigger infrastructure, going V2P for that.
Customers will continue to see exciting innovations from us, which will make virtual infrastructure practical to use in a disaster recovery scenario. We’re trying to generalize that practicability, so it starts to leverage what PowerP2V really delivers.
We increasingly see it as a platform of technology that brings portability to the operating environments in your data center. It’s another way of saying it’s the complement of utility computing. Where utility computing is often described as a way to bring more flexibility to how resources are made available to the applications, in effect move the resources to the application, the technology we’ve built around PowerP2V makes it possible to move the applications to the resources.
If you put the two together, you get the ideal pragmatic way of managing the data center. I think you need a mix in today’s economy. With what PlateSpin is delivering and building for the future, at any given time you could take a server in your data center and move it to where the resources happen to be in order to provide the service levels required. It’s not there to do real-time migrations, like covering a spike at a given hour with more memory; that’s not a reason to migrate your server. But looking at a group of servers, you may want to create a different mix of how they relate to the hardware, using the technology that we have.
One of the other innovations we previewed at the VMworld show was some technology that would reside above the OS portability platform. It would analyze the data center and in effect provide continual optimization. It’s a complement to the load balancing, scale up-scale down instant solutions that people are building that deal with the spikes, because it will look over longer periods of time and provide the best match between the resources and the applications so they can run in a stable mode for the longest period of time.
One example is if you have two virtual machines on a server, and one spikes every Tuesday, Thursday, Saturday and the other spikes every Monday, Wednesday, Friday then the analytics would say that those two can co-exist properly on a machine forever. Therefore one of the things our software will do is identify those cases and migrate your servers to that configuration as it best can determine.
As they scale up and scale down, it may over time recommend other movements of servers in the data center, which can be triggered in an automated way or a scheduled way to move them around. It will take into account blade computing, different forms of virtualization, physical computing models, look at service level delivery over time and start to match things up properly.
We were demonstrating that at VMworld, including a new user paradigm, which had a recommendation engine built into it. As the system detects that things are not where they should be, it will start either making recommendations to the IT person, or it will automate moving them on its own using maintenance windows of the period to do the migrations.
VSM: Will there be a reporting component built into that?
SP: Absolutely. There are a number of new products we’re planning for next year that take all of our technology and create the vertical and horizontal packaging needed for different business needs, whether it’s Disaster Recovery, server optimization, consolidation or business service provisioning.
VSM: How is the product licensed?
SP: We have three licensing models, the starter pack, the one year unlimited license for the data center, and the perpetual license for the data center. We also have an enterprise license that we do on a case-by-case basis. The one year unlimited and perpetual licenses are equally popular, and the starter pack is most popular as the lowest-priced alternative.
Newer accounts often begin with the starter pack, then upgrade to the one year unlimited or perpetual license as their project matures.
VSM: Once people get their hands on the starter pack, they definitely want more.
SP: Absolutely. We just launched the PowerP2V product at the beginning of June, so we’re still only a few months into its lifecycle. We have a lot of interesting customer relationships now, both small and large organizations. We’ve made it easy for people to accelerate the projects they’re implementing around virtual infrastructure by matching up the 25-pack with their own program. So if they’re getting two servers for virtual infrastructure, putting ten VMs on each, the starter pack is a good match for that, instead of forcing them to think about unlimited or perpetual licenses at the early stages. Then as that project grows, and the success of PowerP2V has been proven, they then upgrade to the one-year or the perpetual license.
We’re also trying to provide easier ways to see what the software does. One of the interesting problems of selling a product like ours is that we don’t really provide a trial scenario. That’s another reason we offer the 25-pack, as a reasonable compromise. And if somebody buys it and we can’t make it work for them they can return it in a reasonable period of time.
VSM: How is the support?
SP: The customer base, from a survey perspective, would say that we provide one of the better support programs that they’ve seen, both in terms of responsiveness and ability to solve problems remotely.
There are a few types of typical issues. The first is problems with installation, which happen only occasionally. PowerP2V is very simple to use and install and requires no pre-requisites not commonly found on Windows servers.
The next level is problems related to hardware. From the beginning we engineered the product to be very quickly upgradable if people come across missing drivers for their equipment. For the first two months we had a fair number of those, maybe one in three, but now we’re down to one in ten or fifteen. When that happens we find out remotely the brand and driver requirement, and we’ll typically turn it around in two business days. We’re very responsive with patches overall to help get the customer going.
One thing we designed into the product is to use Windows driver models for migrating Windows machines, and Linux driver models for migrating Linux machines. So we match up against what the customer has a lot better than a solution based entirely on a Linux pre-boot environment.
The third area of support has to do with individual cases, servers with non-standard configurations. Those we have to deal with on a case-by-case basis, so we built into the product the notion of sending a support request from the software itself directly to our support desk. It packages up the inventory and configuration and all the details so the support people can determine if it’s a configuration we support or not, and work through the issues as appropriate. Those requests have gone down quite a bit as well after the first few months. Now largely we’re staying in step with the typical calls you get from new customers, so at any given time there are five or eight customers active on the support list.
VSM: One of your marketing points is that the product is easy to use, so you don’t have to be a virtualization expert. And that brings its’ own set of support issues.
SP: That’s true. The only real expertise people typically need is that they need to know networking, because of the server-to-server communication. If you’re not familiar with networking or operating systems, you’ll have some issues with any solution like ours, whether it’s from PlateSpin or from other vendors.
But the operator time, and how long it takes you to learn how to use the product, can be measured in minutes. We’ve had many people confirm that. It’s a very simple product, and they like the interface style we have, the approach we take, the flexibility and the detail they’re provided when the conversion job is running.
The demo can be very impressive. It’s a good example of taking what is otherwise a very complicated manual task and making it really simple. It certainly is simple when it all works, but it does have a lot of moving pieces to it, so there are a lot of real world variables that can get in the way. When it works, it’s very easy to use. Invest 2-3 minutes of your time, come back an hour later and your server’s running in a virtual machine.
VSM: Do the hardware requirements cover a broad spectrum of manufacturers?
SP: Yes, we made an effort to cover HP, IBM, Dell and other vendors and many of the hardware variations that they have. The support issues come down to mass storage devices and NIC cards. That’s really where the variability exists, so we have support for hundreds of them in our checklist.
The software won’t start the conversion if it thinks it can’t finish the conversion, so there is a lot of pre-flight checking that goes on within the software. It will tell the user that it can or can’t do the conversion before you start.
VSM: What’s next?
SP: At VMworld one of the things that the keynote speakers were talking about is that VM technology will become so prevalent that in about six years 70-80% of all servers will be deployed in virtual machines of some form or another. And not too long after that it will be 100%.
Whether it’s VMware technology or their technology and a combination of others, it’s still a pretty impressive transformation for the data center. One of the things we think is essential to its success is how you can deal with the migration issues that come into play. Given that the typical data center is always a mix of infrastructures, the place that we’ve positioned PowerP2V and all of its successor products is very squarely trying to address that need and be the leading provider of flexible and highly automated migration solutions.
If you turn it back on the IT industry, virtualization really goes against the notion that my hardware product can be different than yours. Virtualization hides the differences. You could argue its succeeding as a business opportunity for VMware to overcome the fact that the hardware vendors and operating system vendors never set up appropriate interoperable standards. The PowerP2V product family effectively removes the impact of hardware dependencies from the entire environment, allowing the IT specialist to move servers around at will.
VSM: Until every box is a white box.
SP: I think the IT specialist would be quite happy to live in that type of world.
*******************************
For more information visit PlateSpin at www.platespin.com.
To see a slide presentation of the PowerP2V event, visit the Events section of the Virtual Strategy Magazine Web site.
|
