virtualization.info

One of the biggest enhancements expected with the next version of VMware Infrastructure (possibly called vSphere 4.0) is the new pluggable virtual infrastructure, which will allow customers to replace the standard VMNet virtual switch with 3rd party software switches.

The first company to offer such product will be Cisco, which announced the Nexus 1000V at VMworld 2008 last September.

After seeing the virtual switch command line for in action and its architectural diagram, we now have extensive details about its features, thanks to an exclusive virtualization.info interview with Paul Fazzone, Product Manager of Nexus 1000V at Cisco.

Fazzone also provided a key information about the release date of the virtual switch: H1 2009.
This date makes very likely that both ESX 4 and Nexus 1000V will be released at the imminent VMworld Europe 2009 in Cannes.

virtualization.info: Why a virtual infrastructure administrator should consider deploying the new Nexus 1000V? Isn’t the VMware virtual networking good enough? What features this new product can bring in?

Paul Fazzone: There are 3 key benefits to the Cisco Nexus 1000V with VN-Link (Virtual Network Link) capabilities. They are as follows:

• Policy-Based VM Connectivity: allows network properties to be defined in the network (including switch, NIC and trunk configs), applied to the VM via Virtual Center and enforced throughout the VM’s lifecycle (even through vMotion).
  
  VI Admin Benefits:
  • Accelerate & Simplify deployment of new ESX hosts
  • Ensure proper connectivity & networking safeguards are in place
• Mobility of Network & Security Properties: VMs need to move (vMotion, DRS, SW Upgrade, etc). The Cisco Nexus 1000V solution provides vMotion for the network properties and connection state ensuring continuous operational insight and security policy enforcement.
  
  VI Admin Benefits:
  • Prevent ESX host/network configuration discrepancies from impacting vMotion
  • Gain consistent visibility into VM-level I/O even during vMotion
  • Secure I/O to VMs located in the DMZ
• Non-Disruptive Operational Model: The Cisco Nexus 1000V brings enterprise and service provider class networking capability to the virtual network access layer without disrupting the virtual infrastructure administrator’s regular workflow.
  
  VI Admin Benefits:
  • VM workflow doesn’t change
  • ESX vSwitch configuration & management responsibility offloaded
  • Data Center operations teams equipped to respond quickly to applications issues

In addition to the key benefits mentioned above, the Cisco Nexus 1000V introduces many new networking and security features to the virtual network access layer. These features include:

• Switching Features
  • L2 Forwarding
  • VLAN Segmentation
  • IEEE 802.1Q VLAN Tagging/Trunking
  • Port state follows VM (Network VMotion)
  • Tx Rate Limiting
  • Rx Rate Limiting
  • NIC Teaming/Port Channels
  • Asymmetric Port Channel
  • Multicast –IGMP Snooping
  • Security Features
  • Private VLAN (Isolated, Community, Promiscuous Trunks)
  • Access Control Lists (ACLs)
  • Port Security
  • DHCP Snooping
  • IP Source Guard
  • Dynamic ARP Inspection
• Management Features
  • Virtual Center Support
  • Standard network command line interface (CLI)
  • Consistent networking features with physical network
  • Port Profiles
  • Port Profiles with Inheritance
  • SPAN – Port Mirroring
  • ERSPAN – Remote Port Mirroring across L3 boundaries
  • Netflow v5
  • Netflow v9
  • XML API
  • SNMP v3 Read/Write
  • CDP v1
  • CDP v2
  • Syslog
• System Features
  • High Availability with Active/Standby Supervisor
  • SSH/Telnet

The Cisco Nexus 1000V enables a network management & operations model at the virtual network access layer that is consistent with the overall data center network infrastructure (physical network switches & routers). This enables customers to more easily integrate virtual machine environments into the existing data center infrastructures.

VI: How the Nexus 1000V will be deployed? Is it a virtual appliance, a core component of the ESX 4 kernel, or a piece of software that must be installed on a dedicated physical server?

PF: There are 2 architectural components to the Cisco Nexus 1000V. The first is the Virtual Ethernet Module (VEM), a Cisco developed offering of the VMware vSwitch. The Cisco Nexus 1000V VEM runs in each ESX host as an embedded kernel module and performs local switching and network services, policy enforcement and statistics gathering for the individual virtual machine interfaces. The VEM operates at the same layer as the VMware vSwitch and is compatible will all servers and NICs listed in the VMware Hardware Compatibility List (HCL).

The second component of the Cisco Nexus 1000V is the Virtual Supervisor Module (VSM), which is the management interface for the solution. The VSM is a version of the Cisco NX-OS data center operating system, which will be available as a virtual appliance and available for download from Cisco.com. The VSM is responsible for network configuration of the VEMs, communication with VMware’s vCenter and management of up to 64 separate ESX hosts (from a networking perspective). The VSM will also be available from Cisco pre-packaged as a physical appliance so that customers have a choice of running it on their own hardware or Cisco supported hardware.

VI: Will it work with ESXi 4.0 or just with the “traditional” version of ESX?

PF: The Cisco Nexus 1000V will work on future versions of both embedded (ESXi) and traditional versions of ESX.

VI: How virtual infrastructure administrators can manage it? Is it going to have just a command line interface (CLI) reachable by TCP/IP? Or is it going to also have a web GUI? Its control panel is going to be integrated with vCenter?

PF: With the Cisco Nexus 1000V, the virtual machine workflow and lifecycle don’t change inside of vCenter. The virtual infrastructure administrator is still in charge of VM creation and management. As part of that VM creation, the VI administrator is still responsible for applying Port Groups to the individual VM VNICs. The main difference the VI administrator will see with the Cisco Nexus 1000V deployed is that the vSwitches, physical NICs and port groups for a given ESX host are automatically configured to work with the upstream network. A library of available port groups is automatically displayed inside of vCenter and the VI administrator can apply them to the VMs to meet their business needs. The application of these port groups automatically triggers the Cisco Nexus 1000V to provide the appropriate network services to a specific VM VNIC. Managed as a traditional network device, the Cisco Nexus 1000V supports the Cisco NX-OS CLI. The Cisco NX-OS operating system also supports a full XML API along with SNMP read/write capabilities so that the Nexus 1000V can be easily integrated with existing data center management solutions.

VI: Is the Nexus 1000V feature-set limited compared to its “physical” counterparts? If so what capabilities are missing?

PF: The feature set for the Cisco Nexus 1000V is very data center focused and has been modeled after products like the Catalyst 4900 and 6500 series switches. The features supported are key requirements for enterprise and service provider customers alike. The feature set has also been enhanced from a traditional “physical” network platform to address key requirements only found in a VM environment like mobility and dynamic provisioning.

VI: How Cisco will be able to guarantee a certain level of performance for this virtual switch? We know that in a virtual infrastructure the performance of any virtual machine can be impacted by the others if there is no resource reservation in place.

PF: The VM kernel has guaranteed CPU and memory available to it. Since the N1K runs in the VM kernel, and the VM kernel controls memory and CPU allocation to the individual VMs, the Cisco Nexus 1000V is safeguarded from being starved from resources. The case is the same with the VMware vSwitch.

VI: How the product will be update? Administrators will be able to use the VMware Update Manager (VUM) or Cisco will provide a new virtual appliance image?

PF: The Cisco Nexus 1000V Virtual Ethernet Module (VEM) inside of the ESX host will be upgradeable using either VMware Update Manager (VUM) or directly through the console. The Cisco Nexus 1000V Virtual Supervisor Module (VSM) will be upgraded like a traditional network device supporting High Availability and Stateful Switchover between 2 VSMs.

VI: Who will sell the product? Cisco or VMware? Will VMware be able to bundle it with VI4? Who will provide the support? Can you give us an idea of the price range and the licensing model?

PF: The product will be sold and supported by Cisco and it’s channel partners. Given the overwhelming interest in the Cisco Nexus 1000V from customers, we are exploring additional channels for the product as well. Information on these additional channels along with pricing and licensing information for the Cisco Nexus 1000V will be announced at product availability.

VI: When the product will be available?

PF: The Cisco Nexus 1000V is targeted for release in the 1 half of 2009 in conjunction with an updated version of VMware’s ESX software.

VI: Is Cisco working with other virtualization vendors (namely Citrix and Microsoft) to bring the Nexus 1000V on other hypervisors?

PF: Cisco does not comment on products that have not been announced. While the Cisco Nexus 1000V architecture has been developed in a hypervisor agnostic manner, the initial product will only offer support for VMware ESX environments.

So far the Xen hypervisor offered a unique opportunity to enter the virtualization market as a major player.
Top industry vendors invested (Citrix, through the XenSource acquisition, Novell, Red Hat) or are investing (Sun) million of dollars to bring it inside the enterprise business.

One the first companies embracing Xen is Virtual Iron which achieved some notable goals in the last few years:

• it accumulated the impressive capital of $65 million in five investment rounds
• it closed major OEM agreements (with HP and IBM, with Dell, with Provision Networks, with Arrow Electronics) and partnerships (with Microsoft, with PlateSpin, with FalconStor, with Reflex Security, with Fabric7)
• it provided a broad range of certified storage (NetApp, IBM, EqualLogic, Data Core) to its customers
• it launched its own channel partner program since more than one year already

and yet the company has to face major compeitition for all the other big players mentioned above.

virtualization.info met Ed Walsh, the new CEO of Virtual Iron and asked questions about the competition, the go to market and acquisition strategies, the technology roadmap, the relationship with Microsoft, and more.

virtualization.info: For its virtualization platform Virtual Iron adopts the Xen hypervisor. Despite the project is open source it was historically influenced by XenSource. What changed since the Citrix acquisition? Is now easier or harder to work on new patches? Is Virtual Iron still has the same commitment on this project?

Ed Walsh: Yes, Virtual Iron is absolutely committed to the Xen hypervisor and the Xen project. There are several factors that make the project stronger than ever:

• The recent changes in the oversight of the project make it more independent and objective. Companies like AMD, HP, IBM, Intel and others are now playing a much stronger role in the project. There are also important new contributors further strengthening the development efforts of the project such as Sun and Oracle.
• At the same time, with Ian Pratt and Keir Fraser leading the project, there is important continuity and stability.
• You also have the natural maturity of the hypervisor itself. This is exactly what open source is about - the power of many. As a result, Xen as a hypervisor, is more proven, more reliable and easier to work with and build on.

VI: The open source world offers an alternative to Xen called KVM, which got a lot of traction in the last year despite it's a young project. What would happen if most of the current Xen contributors would switch to KVM? Is Virtual Iron able to sustain the Xen development alone?

EW: The hypervisor is a very small component of the Virtual Iron platform. And which hypervisor we use is largely irrelevant for our SME customer base. They are only concerned with whether the product works. Whether we use Xen or KVM is not important to them. It is important to Virtual Iron obviously and we feel Xen is more mature and offers us and our customers the best performance.

If KVM ultimately developed into a better alternative, and it provided real advantages for our customers, than we would be open to integrating it into our platform. We are still the only company that has a history of integrating with more than one hypervisor. We initially developed our solution based on our own hypervisor before deciding to migrate to the Xen platform to leverage the power of the Open Source community.

VI: Some of the newest players in the virtualization market, like Oracle and Sun, decided to adopt Xen as well. They will join Citrix, Novell, Red Hat and obviously Virtual Iron. What is the company long term strategy to differentiate from all these competitors?

EW: I think there is a large misperception in the market regarding Xen and the notion that all Xen solutions are somehow the same. They are not. We all leverage the same hypervisor, but we all build unique value on top of it. We also target very different types of customers and market segments with the way we package and price our unique offerings. At the end of the day, it's about delivering real solutions for target customers.

Virtual Iron is gaining great traction amongst SME customers because these buyers demand a full-featured solution that can support all primary use cases - without all the cost and complexity that they find in VMware. Virtual Iron is the only solution on the market that can deliver both. This segment of the market also holds the greatest potential for growth in our view. And I think our growth trajectory, which far outpaces where VMware was at the same time in its history, validates the size of this opportunity.

Oracle and Sun are targeting entirely different market segments, as are Red Hat and Linux. Whether they use Xen is irrelevant to most users.

Our chief strategy officer, Tony Asaro, commented on this whole issue in detail recently on his blog, titled Transcending Xen. I'd encourage you and your readers to see what he had to say.

VI: Just like any other virtualization vendor out there, Virtual Iron suffers the limitations imposed by ISVs which don't support their applications inside virtual machines.
The most striking example comes from Oracle, which refuses to support its database server on any hypervisor but its own. What do you say to your customers that adopt Oracle?

EW: Per my comment above, I think Oracle is targeting those companies's running its full application stack. In general, these are not the types of SME customers Virtual Iron is targeting, although we do have customers running their Oracle database on Virtual Iron. Honestly though, it's really a non-issue for us. It may be an issue for VMware, because they target the same type of customer that Oracle is targeting with this policy.
Ultimately though I think the market will decide. This is the beauty of the free market.

VI: Recently Virtual Iron closed its 5th round of investments, reaching a remarkable equity capital of $65 million. Despite that the company decided to not exhibit at the first European edition of the biggest industry event we have today for virtualization: the VMware VMworld. What's happening? Isn't Virtual Iron interested in the European market?

EW: Virtual Iron is absolutely committed to the European and Asia Pacific markets, which account for 40 percent of our business. We recently expanded our distribution channels in Europe via agreements with Avnet and TechData, and have done the same in Asia Pacific with Avnet. We also have significantly increased our investment in these markets including online marketing, channel programs and new resources on the ground.

Virtual Iron was in attendance at VMworld Europe and we chose not to exhibit because the event is owned, operated and controlled by VMware. We felt it would be very difficult to get our message across effectively in that environment. VMware owns the stage and everything that happens from the time the show begins to the time it ends. In addition, the sponsorship fees were hefty and we felt it was a bad marketing investment. We've learned this from our participation the past two years at VMworld in the U.S. and we saw it again first hand in Cannes where VMware seemed to tighten their grip over the proceedings even more. Our marketing efforts are better leveraged in other areas. We do not plan on exhibiting at the next VMworld event in the U.S. either.

VI: How Virtual Iron will invest the money collected so far? Are there acquisitions at the horizon?

EW: Virtual Iron has always been committed to delivering leading-edge server virtualization solutions for the mainstream market and we will continue to invest in product development. We have also taken a much more aggressive marketing and channel development stance and are investing much more in these areas. This includes significantly increased investment in Europe and Asia-Pac.

We are not currently looking to acquire other companies. We feel our current product more than meets the requirements of our client set. Of course, we continue to have an aggressive product roadmap.

VI: In August 2007 Virtual Iron joined the Microsoft Interop Alliance. Which kind of benefits that partnership is bringing to your customers besides a seamless import of Virtual Server 2005 virtual machines? How Virtual Iron will implement the interoperability with upcoming Microsoft Hyper-V?

EW: Our participation in the Microsoft Interop Alliance is focused on enhancing interoperability for our users with Microsoft systems. As you would expect given our SME focus, the majority of our customers are running Windows-based infrastructures.

In November, we also announced our plans to join the Microsoft Virtualization Validation program. This will ensure complete Microsoft support to Virtual Iron customers running Microsoft Windows Server as guest operating systems in Virtual Iron virtual environments. Our customers will have greater access to Microsoft's comprehensive support capabilities and it will improve support hand-offs between the two companies, increasing satisfaction and adoption. The complete text is available here.

VI: Recently the virtualization industry seemed to evolve into three main directions: VDI, disaster recovery and data center automation. We already know about the Virtual Iron efforts in the VDI space through the partnership with Provision Networks (now acquired by Quest). What the company is doing or will do in the other two market segments?

EW: Virtual Iron is currently delivering solutions in all three of these areas.

Regarding data center automation and disaster recovery, it's important to point out that policy-based management has been a core competency for Virtual Iron since Version 1 of our product. Today more than 50 percent of our customers today are using virtual infrastructure capabilities like policy-based management, transparent workload migration and automated failover for disaster recovery. Virtualization is also a huge enabler for disaster recovery in SMEs. Traditional DR strategies simply aren't an option for these companies. It just isn't feasible for an SME. You can also expect to see us secure additional partnerships in the disaster recovery area.

These are some of the reasons customers choose Virtual Iron. It has all the advanced capabilities for data center automation and advanced use cases like disaster recovery without all the cost and complexity. We have highlighted some of our customer implementations here in a new Blog called Customer Spotlight. I think these examples highlight the types of advanced functionality our typical SME customers are using with Virtual Iron.

VI: Another application of virtualization that several vendors tried to push in the last two years is the so called Virtual Appliance. Virtual Iron started to endorse this approach in January 2007 but it doesn't seem it got any traction so far. Why the market answered so tepidly?

EW: Virtual Appliances are still a very new concept. They make great sense on paper and present an exciting opportunity, but the packaging, pricing, sales, implementation and support systems that will make them common practice in the data center are not there yet. There are still no common standards for virtual appliances. Microsoft is still charging for Windows guests running in production within virtual appliances.
Most ISVs have yet to embrace virtual appliances. The industry has not even really figured out how to sell virtual appliances yet. Take something as fundamental as the channel (the primary distribution vehicle for the server virtualization market). How do they buy, sell and license a virtual appliance? There are a whole set of fundamental pieces like this that are not in place yet. I think the market will ultimately embrace virtual appliances, but it's a matter of time before these other pieces of the puzzle are in place.

VI: Recently VMware suffered a bad stock performance. This could be a signal that investors don't trust virtualization anymore. What's your opinion?

EW: The virtualization market is very strong, and we believe immune from any broader economic downturn. Some would even argue that virtualization will rise in importance if the business climate continues to slide. I also think SMEs will continue to invest in virtualization regardless of the tech economy. History has shown SME technology purchasing s to be far less impacted by economic factors than large enterprises. That bodes well for disruptive companies like Virtual Iron.

Regarding VMware's recent stock performance, I don't believe that it truly reflects the company's performance. VMware grew 80% quarter over quarter and was immensely profitable. The stock was just overhyped during its IPO, and the last quarter was a reflection of the market's overreaction to some very lofty expectations that were set over the summer. The virtualization space is early enough and large enough for VMware and several other industry players to grow and be very successful.

Quest involvement in virtualization has increased substantially over the last two years. The company initially made a controlling investment in Vizioncore in 2005, acquired Invirtus in June 2007, then Provision Networks in November 2007. Last month Quest finally completed the acquisition of Vizioncore.

While the acquired companies are focused on different virtualization segments (disaster recovery, VDI, P2V migration and virtual machines performance optimization), Quest itself has long been well-known for its focus on software management of applications, databases and operating systems like Windows.

So while Quest is becoming a major virtualization player, its parallel business focus, its acquisition strategy and its current relationship with its subsidiaries in virtualization can be confusing to customers. virtualization.info met Scott Herold, the new Lead Architect of Quest's Virtualization Business Unit, to ask for clarification about the big picture.

virtualization.info: So far Vizioncore, Invirtus and Provision Networks operate as independent subsidiaries of Quest. Are they going to continue this way or they will be merged over time? In both cases, will "core" Quest continue focusing on its current software management business?

Scott Herold: As of right now, each of these acquisitions operates and is managed independently with guidance from the newly formed Virtualization Business Unit within Quest Software. Quest formed this new Virtualization Business Unit so the company can put a primary focus on one of the hottest datacenter technologies in history without detracting from its core Database, Application, and Windows Management business units. You may have already noticed a trend that will continue -- a cross pollination of intellectual property from these various acquisitions that will help strengthen all of Quest's virtualization offerings.

VI: In July, immediately after its acquisition, Invirtus product line was rebranded as part of the Vizioncore offering. Since that time Invirtus didn't release any new product. Which plans does Quest have for this subsidiary?

SH: Invirtus was rebranded as a part of Vizioncore to prevent confusion and challenges in building parallel sales channels within the same overall organization. The first step of the integration process was to rename and re-release all Invirtus products under the Vizioncore name and get the new technologies properly introduced into the channel. Invirtus has a solid set of technologies that, although they weren't released as Invirtus products, still made their way to market. You may notice that right around the announcement of Quest's acquisition of Invirtus, Vizioncore announced P2V Disaster Recovery for vRanger Pro, which makes use of Invirtus' vConverter technology. Invirtus is still working extremely hard as an R&D powerhouse, with the launch of vConverter Pro at VMworld Europe this year as evidence. The product provides the fastest, easiest, and most reliable P2V conversions on the market. Another product of note is vOptimizer Pro which will optimize the storage of ESX Server virtual machines.

VI: Quest has demonstrated substantial in-house expertise surrounding all facets of Windows management, and this may boost the current virtualization offerings or even enable the birth of new products. What should we expect in the coming months?

SH: One of the most daunting tasks on my plate is eventually reviewing more than 120 products (including the Windows management products you mention) and technologies that Quest Software either owns or has invested in to look at them from a virtualization perspective with the goal of either breathing new life into aging products or assisting product managers in determining how their existing offerings can be enhanced and extended into the virtualization space.

VI: Quest has long had a strong focus on Microsoft technologies, and the company has officially announced its support for the upcoming hypervisor Hyper-V. At the same time one of Quest's subsidiaries, Vizioncore, is an historical VMware partner and has never released products for other virtualization platforms. Now that the acquisition is complete can we expect to see any Vizioncore solution for Microsoft Hyper-V?

SH: Part of the overall Quest Software strategy across all organizations is to work closely with our customers to determine which alternate platforms are being seriously looked at, and which technologies would benefit those platforms the most. I can easily say we are not ruling out any particular platform, but ultimately, it will be our customer base that helps us decide what we develop and when.

VI: Quest enters the already crowded VDI market thanks to the acquisition of Provision Networks, which offers a popular connection broker for multiple virtualization platforms. How does the company plan to compete against established players like VMware and Citrix, since they can offer a complete VDI solution on their own?

SH: In your question, you answered part of the question for me. Those that are familiar with Provision Networks Virtual Access Suite know that it already supports all major virtualization platforms. Within Quest Software and Provision Networks we don't see VDI as the ultimate target for this technology. We see Application Delivery as our goal. VDI is one method for delivering applications or desktops to an end user; physical blade servers are another, and application virtualization is yet another. Provision Networks is taking the Switzerland approach to application delivery by supporting all virtualization platforms, as well as multiple other delivery technologies that enhance the offerings of all technology partners involved.

VI: Besides VDI another emerging market for virtualization seems to be the so called VM Lifecycle Management. Quest has a notable experience in Microsoft PowerShell technologies, which could be manipulated to create a powerful VM lifecycle management product for Hyper-V and ESX Server (since VMware is integrating the technology inside VirtualCenter). Does Quest have any plan to enter this market as well?

SH: While you are correct that Quest has demonstrated thought leadership in the PowerShell space and accurate in assuming that the company is always looking out for new ways to add value on top of the Microsoft platform, we do not have any specific development efforts to announce in the space you mention at this point in time.

VI: Another major vendor entering the virtualization market is Sun, with the hypervisor xVM Server and the management solution Ops Center. Quest has announced its support for these two products. In which way will the company support Sun offering?

SH: Initial xVM support will come natively based on work that we will be doing for integration with the other Xen-based platforms on the market. Each implementation of Xen, whether it is Virtual Iron, Citrix, Sun, or Oracle, will require unique customizations that will further enhance the offerings for the particular platform. What we are seeing is that Sun has a strong footprint in the banking and financial industries, and there is a very strong interest in virtualization. Our research in this area has gone a long way in helping us properly prioritize xVM support. Something that is unique to Sun and Microsoft, is that within Quest we have some of the best technical resources available that are more than willing to assist us in providing the proper features for virtualization within these platforms.

VI: Scott so far you covered the roles of Director of Research & Development at Vizioncore and of Director of Engineering at Invirtus. And even before these experiences you already were a popular and appreciated virtualization architect. In your opinion what is the today's biggest challenge in virtualization technologies adoption? Is Quest going to address it in any way in the near future?

SH: I personally still think that the biggest challenges in virtualization adoption aren't technical, but more business process oriented. Even with the successful implementation of virtualization at many large organizations, people are struggling with changing the way they have traditionally run IT as a business. Virtualization brings challenges because new servers are a mouse click away, virtual machines are portable and Disaster Recovery is as simple as copying a few files … and all of a sudden you have a way to track system utilization to levels that haven't been possible outside the mainframe world. Many organizations, without the proper help, continue to treat and manage virtual infrastructures the same way they have for years and simply cannot see all of the available benefits to them.

VI: Do you think that the customers demand for virtualization changed over time? Are companies still looking for the same features? Are they still using virtualization technologies mainly for server consolidation?

SH: The demand has definitely changed over time. Virtualization originally provided a way to consolidate the number of servers running in an environment. Not long after the first few large implementations, people started to notice how much more recoverable virtual machines were and how their disaster recovery plans could be significantly enhanced. Now, the next big thing is changing the way desktops and applications are delivered to the end user. The use cases for virtualization are seemingly endless and moving rapidly.

I think an initial driver for virtualization is still server consolidation, especially in parts of Europe where real estate and energy costs are at a premium. Green computing is not only a popular buzzword; it is a reality, and many organizations are willing to do their part. Once virtualization is introduced, the possibilities are almost endless.

VI: What is the next big thing in virtualization according to the Quest vision?

SH: As I mentioned earlier, there has been a lot of acquisition activity of late for Quest in the virtualization space. Now the company is focused on turning this activity into customer benefit and company revenue. There's no question that virtualization is the hottest space in IT in 2008, and as the leading provider of IT management tools, products and solutions, Quest is now naturally nurturing a leadership position in managing virtual environments. This is a logical extension from Quest's existing core business, supporting a comprehensive portfolio of management offerings across all tiers of the IT hierarchy. And it's good for Quest's virtualization subsidiaries as well because it gives them broader access to Quest's enterprise customers, an expanded global reach and a greater partner network to tap into than they had before being acquired. All these benefits are coming together in a product roadmap for virtualization that will build from Vizioncore's existing leadership in key customer need areas like business continuity and server consolidation to the ever-growing set of benefits organizations can gain from virtualizing their IT infrastructure.

Sun is the latest major player entering the virtualization space and challenging VMware after Microsoft, Virtual Iron, Novell, Red Hat, Citrix and Parallels. Anyway the company is in a unique position compared to all these competitors.

Sun in fact is able to provide the certified hardware (both servers and storage), two different virtualization platforms (both xVM Server and Solaris Containers), a management solution (Ops Center), and even a VDI connection broker (Sun VDI).
On top of that Sun has a partnership in place with Microsoft to grant interoperability and high performance for Windows virtual machines.

For all these reasons Sun entrance in the market is expected with much interested.

virtualization.info met Steve Wilson, Vice President of xVM, asked ten questions about the overall Sun virtualization strategy and discovered some major news:

virtualization.info: During Jonathan Schwartz's keynote at Oracle OpenWorld, Rich Green, EVP of Software of Sun Microsystems mentioned that xVM Server uses a Solaris kernel in lieu of the Linux-kernel that's typically used in Xen-based hypervisors. What does this mean exactly? How does this change impact the hypervisor's capabilities?

Steve Wilson: Every hypervisor has a control domain that manages the hardware and I/O interactions for the guest operating systems. Hypervisors based on the Xen open source project call this Dom0. For Sun xVM Server, we have inserted a number of datacenter-grade features borrowed from Solaris into the Control Domain that give xVM Server a set of highly unique attributes.
One example is Predictive Self Healing, which allows xVM Server to shield guest operating systems (including Windows, Linux and Solaris) from various types of hardware faults, such as CPU and memory failures. This can lead to dramatic improvements in availability. In addition, guest operating systems (including Windows and Linux) will be able to transparently access ZFS file systems. This opens up a number of interesting possibilities in I/O performance, snapshotting and backup, and even encryption and security.
Another example is the ability to leverage the new Crossbow I/O virtualization technologies to do bandwidth management between guests to ensure low priority guest tasks don't hog all the bandwidth available through the host's physical NIC.

VI: Which other customizations has Sun implemented around Xen in xVM Server?

SW: A key theme for us is interoperability. We're adding the ability for xVM Server to run VM files created for VMware's ESX Server or Microsoft's Hyper-V without modification. We think this will be key to allowing customers to easily adopt our technologies.
Another key theme is management. xVM Server will include a very simple to use, web-based control application. However, for operating at scale, we have an additional product called xVM Ops Center. xVM Ops Center is designed to manage up to thousands of servers (physical and virtual). It includes a simple, but scalable user interface that enables a typical team of administrators to dramatically increase their efficiency in managing a datacenter.

VI: Xen is used as the foundation for several other virtualization platforms (i.e.: Citrix, Virtual Iron, Novell and Red Hat). How does xVM Server differ from these other implementations?

SW: Most customers aren't concerned with whether a given solution is using a Xen hypervisor or not. Rather, they're looking for the best solution, so I really view our competitors as everyone you mention, as well as VMware and Microsoft. However, even viewed against all these competitors, I think Sun brings a lot of unique value.
Sun xVM is a virtualization and management platform designed to run in the heart of the datacenter. With the unique capabilities of xVM Server and xVM Ops Center, customers will be able to see improvements in availability, scalability and manageability.
One other differentiator we shouldn't overlook is the openness of Sun's solution. Obviously, many players are particpating in the Xen open source community for technology around the hypervisor. However, most of the other players are keeping their management solutions closed, and that threatens to lock in customers. Sun has committed to making xVM Ops Center available freely under the very liberal Gnu Public License (GPL) version 3. Customers like this openness, and partners are particularly excited about the potential for them to get really involved in the development and future direction of the product line. We're now focused on ramping up the activity around the source code base.

VI: One of the major shortcomings of current virtualization platforms is the lack of native clustering/failover capabilities at the host level. What is Sun's answer to this problem? Will xVM Server come with its own HA technology? Will xVM Server support open source native clustering/failover capabilities or will Sun wait for 3rd party commercial products?

SW: Sun already has one of the most popular and mature clustering offerings in the market in Solaris Cluster. We're planning to provide the ability to use Solaris Cluster together with xVM Server for applications where true clustering is required. The integration of Sun Cluster and xVM is an open source project, and we haven't announced commercial availability of this yet, but interested parties can get involved by going here.

VI: Another growing issue has to do with the patch management strategy for hypervisors. How does xVM Server address patch management?

SW: The way we see it, there are actually two issues: patch management for the hypervisor and patch management for the guest operating systems. Both of these are key issues. For small installations, xVM Server includes a simple to use, self-patching system that can automatically download and install the newest patches. We will even leverage ZFS to automatically snapshot the host hypervisor before it patches itself for easy, automatic rollback if there are any problems with the patched hypervisor.
However, I think the more interesting problem is the patching and updating of guest operating systems and their associated applications. In 2006, Sun acquired a small company called Aduva that specialized in patching and updating. The updating technology from Aduva has been integrated into the Sun Connection product that is now used by about 200 of Sun's top customers to patch and update up to 3,000 servers at a given customer.
That Aduva technology is now also part of xVM Ops Center, and it gives us the ability to do sophisticated software lifecycle management that includes patch simulation and rollback, automation and scheduling, compliance reporting.
Ops Center 1.0 includes support for patching of Solaris (x86 and SPARC) as well as several versions of Redhat and SuSE linux. Windows patching support will be added in a future revision. We view this area as a place where we can bring customers a huge amount of value.

VI: Will you integrate xVM Server with the recently released Sun VDI? If so, when we should we expect such integration?

SW: Yes. Stay tuned for more information on plans in this area.

VI: Which kind of virtualization-specific operations will be available with the first version of xVM Ops Center? Can we expect things like virtual machines live backup, live migrations (what VMware calls VMotion), automated provisioning, and/or other features?

SW: The first commercial release of xVM Ops Center will ship in the next few weeks - actually ahead of the first commercial release of xVM Server. xVM Ops Center 1.0 is focused on datacenter automation and includes features such as:

• Server discovery and inventory management
• Server firmware analysis and provisioning
• Bare metal Server provisioning
• Patch management
• Monitoring

We're planning the first commercial release of xVM Server, and an update to xVM Ops Center to go with it, for Q2 of calendar year 2008. In that release, we'll be adding a number of specific features to manage virtualized environments, including:

• Full virtual guest lifecycle management
• Management of the domain 0 instance
• Monitoring, management and provisioning of Windows, Linux and Solaris guests
• Migration capabilities (Live, Regular and Cold)
• Simple single host management through direct browser access, as well as large scale multi-node management via xVM Ops Center
• Expansive resource monitoring and analysis
• Guest image storage library management
• Virtual and resource pooling
• Network virtualization and bandwidth management

VI: xVM Ops Center integrates the open source library libvirt which allows cross-platform virtualization management. Will you provide additional APIs for integration with those management products which already handle multiple hypervisors but don't implement libvirt?

SW: Both xVM Server and xVM Ops Center will expose API sets through WS-Management. We'll be putting specs out for this over at http://openxvm.org shortly.

VI: Which kind of benefits will customers experience from the Sun partnerships with Microsoft and Red Hat? Can we expect better performances for Windows virtual machines? What else?

SW: While we are doing a number of things to ensure best-of-breed performance for Windows with xVM Server, it isn't the focus of our partnerships with Microsoft and Red Hat. Our partnerships with these vendors were put in place to ensure that customers will see maximum compatibility in a totally supported environment. Customers shouldn't have to worry if their applications are supported in a given virtual environment. By ensuring this type of rigorous cross-certification, we can ensure that customers are fully supported by both Sun, and their operating system vendor of choice (be that Sun, Microsoft or Red Hat).

VI: How do you plan to migrate virtual machines from 3rd party hypervisors to xVM Server? Can we expect a P2V/V2V migration tool?

SW: We've decided to avoid creating any new file formats for virtual machines, and instead will be directly supporting the native formats of both VMware's ESX Server and Microsoft's Hyper-V. If you have VMs today, they should run unmodified inside xVM Server.

Biggest news after VMware IPO surely is XenSource acquisition by Citrix.

Leading one of the most popular open source projects ever, Xen, XenSource is a critical company to influence (in positive or negative ways) the free world community and the virtualization industry.

virtualization.info wanted to investigate how this acquisition may impact the market, and asked Simon Crosby, CTO at XenSource, and Wes Wasson, Corporate Vice President of WorldWide Marketing at Citrix, to answer ten questions about different topics, including Xen forking, commercial offering changes, partnerships agreements and support policy with VMware.

virtualization.info: Many Xen developers are currently employed by XenSource. How this acquisition will impact future development of open source project in detail?

Simon Crosby: We expect this acquisition to increase development resources dedicated to the Xen open source project, both from Citrix and from the larger community. Citrix believes strongly in the open source model XenSource nurtured so successfully in the past and will do everything in its power to preserve and grow it.
We fully support the efforts already underway at XenSource to establish a more independent oversight of the project to ensure that it remains strong, innovative and focused on its core mission. At the same time, we intend to aggressively fund our own contributions to the Xen open source project in a way that benefits the entire community.

VI: What will happen to vendors currently using Xen hypervisor as base for their products (namely Virtual Iron, Novell, Red Hat, Amazon and Sun)? They will be able to continue doing same thing without any impact?

SC: This acquisition is great news for vendors using the Xen hypervisor for their products. The will benefit from increased resources dedicated to the open source Xen project from Citrix, more mature processes driving releases, broader industry participation and greater industry visibility for the Xen virtualization engine.

VI: Is it technically possible for one of these vendors to start a fork of Xen?

SC: As with any open source project, it is always technically possible to take the freely available source code and fork development. Vendors who chose this route to create their own proprietary hypervisor will not, however, be able to call it Xen as it would no longer comply with the open compatibility requirements that are core to the Xen open source vision.
We doubt many vendors will chose this route as there would be virtually no upside in doing so. In addition to losing access to the Xen brand, any vendor who chose to fork the code would lose access to more than 250 of the world's best virtualization developers and the tremendous testing that the community does against all major hardware and software platforms.
There is no proprietary advantage to be gained by trying to outdo the community on the underlying engine. The real value is in taking advantage of the open source engine and building innovation and commercial value on top of it. This is where we expect most vendors to focus their energies.

VI: What will happen to existing technology partnerships that XenSource has in place with Microsoft and Symantec?

SC: The existing partnerships with Microsoft and Symantec will remain in place and will actually be strengthened by this acquisition.
Microsoft has publicly stated its strong support for this acquisition with combines XenSource with Citrix, one of its strongest partners and the only two time winner of Microsoft's ISV of the Year award. Citrix has a long history of embracing the Microsoft Windows platform and adding value on top of it and will do the same with the upcoming Viridian platform. Citrix will also work closely with Symantec to build and extend the existing XenSource relationship.
The companies share a common belief that most customers would prefer to manage storage in their virtual environments with the same solutions they use to manage their physical environments. The existing partnership with Symantec is a perfect example of this philosophy.

VI: Will Citrix continue to develop XenEnterprise with same focus on Linux? Or company will shift its efforts in improving Windows guest operating systems?

Wes Wasson: Windows is the most widely deployed operating system in the market today on both the server and desktop side. As a result, XenSource has focused very heavily on supporting this market, a focus which will not change with this acquisition.
Broadly speaking, however, Citrix views XenEnterprise as the best virtualization platform for ANY guest operating system and will continue working hard to ensure that it remains the best platform for supporting both Windows and Linux guests.

VI: Will Citrix acquisition impact on current XenSource prices?

WW: Citrix does not anticipate any price changes as a result of this acquisition.

VI: Will Citrix continue to release the free edition of XenEnterprise called XenExpress?

WW: Absolutely. We believe XenExpress provides a great platform for single-server virtualization and will give our customers and partners an easy way to try the Xen experience in their environments.

VI: Is Citrix planning to integrate all existing virtualization technologies in a single platform or they will continue to appear as separate solutions?

WW: We will actually do both. The current XenSource portfolio will continue to exist and to expand in the fast-growing server virtualization market. In addition, the XenEnterprise platform will be integrated into other product lines within the Citrix application delivery infrastructure portfolio, creating additional value add for those customers.

VI: With the XenSource acquisition Citrix is able to offer hardware virtualization along with presentation virtualization and application virtualization. Doesn't this overlap the Microsoft offering?

WW: This is essentially the same question Citrix has faced for more than 15 years. During that time, we have managed to build a highly successful $1 billion industry leader based on a strategy of embracing and extending the Microsoft platform. Our successful relationship with Microsoft is successful because of our fundamental belief in Windows as an innovation platform. By working closely with Microsoft, Citrix has been able to take advantage of underlying capabilities of the Windows platform and add value on top of it, a strategy which creates a positive business model for Citrix and drives considerable revenue for Microsoft.
This strategy is very different from the strategy espoused by VMware, who sees itself as another operating system vendor and has publicly identified Microsoft as its top competitor. In sharp contrast, Citrix believes our acquisition of XenSource will extend our longstanding relationship with Microsoft into additional virtualization arenas, allowing us to add value on top of Viridian, a strategy we believe will be extremely positive for both Citrix and Microsoft. By ensuring broad interoperability and compatibility between the Xen virtualization engine and the upcoming Viridian platform, customers will be able to freely deploy on Xen-based products today, knowing they can easily migrate to Viridian-based solutions in the future. Both Citrix and XenSource believe virtualization should be a core part of the Windows operating system.
XenSource does not get any revenue from the underlying hypervisor, nor will Citrix. Our revenue will come from adding a compelling set of application delivery value propositions on top of the hypervisor, whether it is Xen or Viridian. The more successful Viridian is, the more opportunity Citrix will have to sell value-added application delivery offerings on top of this platform, much as we do today with Windows Server in the Terminal Services market.

VI: Now that Citrix becomes a virtualization player, what will happen to its VMware support policy? Will Citrix continue to support VMware based VMs in Citrix Desktop Server and the desktop broker feature of Citrix Presentation Server?

WW: This acquisition does not change anything with regards to our support for customers using VMware. Citrix has always believed in broad support for whatever solutions customers currently have deployed, and that stance will not change in any way. Our mission is to help customers deliver applications and desktops to their end users, regardless of the source of those resources.
While Citrix and VMware will compete in some areas as a result of this acquisition, we will also need to work together in many areas to support our common customers. Customers choosing Citrix can rest assured that our philosophy of openness and support for solutions from other vendors will continue.

IBM pioneered virtualization 40 years ago with mainframes. Since that time the technology has been ported to x86 market, mainly thanks to the effort of VMware.

Now that virtual machines are revolutioning the whole IT industry, IBM is demonstrating its experience putting huge efforts in hypervisors (Virtualization Engine and Xen), management tools (Virtualization Manager and Tivoli Usage and Accounting Manager), desktop solutions (Virtualized Hosted Client Infrastructure), and benchmarking (vConsolidate) development.

virtualization.info met Rob Sauerwalt, Global Brand Manager for IBM System x, to discuss about current company involvement in virtualization and his point of view about raising market trends like virtual appliances and application virtualization in an exclusive interview.

virtualization.info: Before the end of the year IBM and Intel announced a new joint initiative to provide customers new tools to effectively deploy virtualization. The first tool announced has been the benchmarking methodology called vConsolidate. When will vConsolidate be available? Will it only be available to customers purchasing IBM systems?

Rob Sauerwalt: The vConsolidate benchmark methodology is available now. As with other benchmarks - such as SPECweb - anyone can replicate/follow the instructions and get a vConsolidate result. Information on vConsolidate is available on Intel's Web site at http://www.intel.com/technology/itj/2006/v10i3/7-benchmarking/6-vconsolidate.htm/

VI: vConsolidate is able to measure performances on non-IBM systems and/or non-Intel hardware?

RS: Yes, the vConsolidate methodology works on hardware from any x86 vendor. It currently does not encompass non x86 systems.

VI: As you know VMware too is working on a benchmarking methodology called VMmark. How does vConsolidate differs from it? Will measurements obtained through VMmark be comparable with ones obtained from vConsolidate?

RS:Not a lot is publicly known about VMmark, as there is limited access to the methodology because of VMware concerns that the source scripts contain intellectual property. We work very closely with Intel to ensure benchmarks are developed to meet the needs of IBM clients.

I recommend you talk to Tom Adelmeyer at Intel who is heading up the vConsolidate methodology, and is likely in the best position to draw comparisons to other methodologies.

VI: In November, 2006 Standard Performance Evaluation Corporation (SPEC) announced it's working on a standard method for virtualization benchmarking. Both IBM and Intel are members of this working group. How does vConsolidate relate to the SPEC effort?

RS: IBM and Intel have discussed vConsolidate with SPEC at a high level and have shared what we learned while developing the vConsolidate methodology. We plan to continue working with SPEC.org on the development of a benchmark that the entire industry can get behind.

VI: IBM also announced the launch of a Virtualization Resource Center (VRC) in early 2007. What exactly is the VRC? Who will have access to it?

RS: The VRC is an IBM program that is jointly sponsored by IBM, Intel and VMware. The program allows customers to test drive IBM hardware that is preconfigured with VMware, virtual machines and management software. The hardware is shipped to customers for free for a 60 day trial and IBM provides technical and set up assistance.

The VRC pilot program is currently accessible to IBM customers in the Americas and we expect to expand it in the future.

VI: Physical servers aimed at running virtualization platforms must be particularly flexible in configuration and highly reliable. Is IBM already addressing these requirements? Or is it preparing some hardware specifically tailored for virtualization?

RS: In the forty years since IBM pioneered virtualization on the mainframe, we have led the industry in extending the benefits of virtualization - including higher availability, higher utilization, shared I/O, memory caching and so on - to other server architectures and platforms. IBM first onto what we now call our System i architecture and then the System p and, finally, our x86 or System x architecture. Enhancing virtualization was one of the founding principles of IBM's X Architecture. Memory is, as your readers probably know, the precious resource in today's virtualized environment. Through the X3 architecture, IBM scalable (4P+) x86 systems have the ability to provide more physical memory than any other Intel-based server offering from a full line supplier (HP, Dell, etc.)

VI: One emerging trend in virtualization is about so-called "virtual appliances." What is IBM's position on the technology? Will we see the company's products being delivered through virtual appliances anytime soon?

RS: The goal of virtual appliances is really to make it easier to exploit virtualization by delivering the virtual machine with the software in a preconfigured/pre-enabled way. IBM is extending a virtual delivery capability to our customers with an offering from our Tivoli portfolio called IBM Tivoli Provisioning Manager. Tivoli Provisioning Manager has been available from IBM for several years and can help deliver some application software through virtual machines. Provisioning Manager was recently updated to include technology from our acquisition of Rembo in 2006 to also push out operating systems to desktops and endpoints. By readily delivering, installing, removing and updating software - both operating systems and application software - managers, depending on how many endpoints software is being installed to, can save minutes, hours, days or even weeks and, because the removal and delivery is almost immediate (it takes only a few seconds) - security can be greatly enhanced as well.

VI: IBM's effort in the open source community, including virtualization technologies like the Xen hypervisor has always been well known. Despite its nature - a large part of Xen developers are employed by a commercial company called XenSource, which signed a partnership deal with Microsoft in August, 2006. What is IBM's perspective on this partnership? Does IBM worry the deal will negatively impact open source virtualization? Did this concern influence IBM's involvement in Xen development?

RS: While Microsoft's engagement in the open source community is commendable, the Microsoft/XenSource partnership has no baring on IBM's contribution to and involvement with Xen, or in any other part of the open source community.

VI: Microsoft announced that it will, a few years from now, release its own hypervisor offering for free in every new copy of Windows Server. Do you think this move will change the virtualization market landscape?

RS: We believe it will positively impact the marketplace. Any time a valuable function, like the Windows Virtual Server, is brought to the market at an attractive price point - especially when it is bundled, for free, into another offering, there will be an increase in adoption. Market activities like this, and like Xen highlight the importance of choice when clients build out a virtualized infrastructure for x86. This is why IBM has built the virtualization management environment to support all choices the client may make. Regardless of what technologies are widely adopted, IBM will be there will to support and optimize their infrastructure.

VI: IBM pioneered virtualization 40 years ago on its mainframes. Since then this technology has evolved, becoming pervasive and taking different forms. So while server virtualization is going more "mainstream", application virtualization is just emerging. In your opinion will customers have to choose between the areas of virtualization or will they coexist? Which trends will we see at that time?

RS: Application virtualization and server virtualization are two trends that are currently advancing independently but will ultimately converge, IBM believes, to the point of mutual dependency. Server virtualization is now, as you suggested, much more mainstream, sometimes even independent of the workload being run on the servers. A better approach to virtualization should be more application dependent - in other words - What is the workload? How much memory is needed? What are the I/O needs?

Application virtualization is perhaps "newer" in terms of awareness but it is making great progress. Think about the phenomenal growth of enterprise grids, the growth of WebSphere XD, Tivoli Workload scheduler and so on. All of these strategies are focused on the workload and moving each workload to the appropriate resource. As architectures like SOA evolve, clients want and, frankly, need to dispatch the right service to the right type of resource, with the needed qualities of service for that workload to get the most from the technology. The "right" resource will often be virtual, no doubt, but putting the right resource at the right time will depend on application virtualization. So, long term, the two virtualization trends will be interdependent.

One possibly major sticking point to the advancement of application virtualization is licensing. IBM announced last summer that we were changing how we licensed and charged for several of our offerings to respond to the changes in how computing is done and several other major and smaller vendors are doing the same thing. Once you remove the traditional mindset that "one server, physical or virtual, equals one license" - you start to see a change in the way the applications and the hardware are used. The sky really is the limit in terms of how far virtualization can go as some of these thoughts emerge in the industry.

Despite the wide and good success the Xen hypervisor project has obtained in the open source community, commercial solutions based on it or incorporating it didn't have similar luck:

• the launch of Novell SUSE Linux Enterprise 10 has been severely judged by Red Hat, accusing the competitor to offer the product while still unsuitable for enterprise deployment
• the interoperability agreement between XenSource and Microsoft has been hit as well by VMware, accusing competitors to take advantage from open source community work without providing anything back
• and finally the launch of XenEnteprise, by XenSource itself, has been compromised by its own CEO's declarations, stating Xen actually isn't enterprise-ready

In this very confused scenario the most direct XenSource competitor, Virtual Iron, which is going to deliver a virtualization product based on Xen as well, decided to talk with virtualization.info: Mike Grandinetti, Chief Marketing Officer, offers company's point of view on Xen maturity, Red Hat involvement in the project, value of XenSource agreement with Microsoft and VMware future.

virtualization.info: Mike first of all thanks for joining virtualization.info and provide readers another perspective on this controversy. Explain Virtual Iron’s level of involvement in the Xen project, and how long the company has been working on the project.

Mike Grandinetti: Virtual Iron has been involved with the open source project for close to a year. We started contributing suggestions and code to the project in early 2006 with the goal of making Xen enterprise ready; i.e. hardening the hypervisor, improving reliability and scalability, increasing I/O efficiency and supporting enterprise-class workloads. Most importantly, we’ve enabled the hardware-assist capabilities provided on the latest generation Intel and AMD platforms which facilitate the support of unmodified 32 and 64 bit Linux and Windows operating systems.

We've contributed and continue to contribute our development efforts to the project. We've also spent a lot of time working with other core members of the Xen community – AMD, HP, IBM, Intel, Novell and Unisys - to understand their interest and needs and to validate our architectural approach and product design. To date, we would estimate that Virtual Iron has invested over 400 man months turning Xen from a project into a product.

VI: Some suppose Red Hat simply wasn’t able to deliver Xen within its own distribution in time for competing with Novell because they didn’t contribute to the project in the time originally planned. Can you tell us how much Red Hat has been active in the development of the hypervisor in the last year?

MG: Like Novell, Red Hat has been involved with the Xen project since its inception 3 years ago. However, it would appear that Red Hat is behind Novell in their Xen development efforts. Like most of the early Xen community, including Virtual Iron, Red Hat believed that the Xen project was more mature and professionally developed than what turned out to be the case.

Red Hat claims to have a small army of developers working on this project. However, it’s unclear to us why they have not been able to make more progress. Integrating Xen into RHEL may have proven to be more difficult than first estimated as they have yet to deliver a single OS paravirtualized on Xen, including their own upcoming RHEL 5 distribution.
Perhaps paravirtualization as an approach is too engineering-intensive in this rapidly evolving virtualization market. For example, with a measurably smaller team, working in a much tighter time frame, Virtual Iron has been able to deliver six unmodified OSs (both Linux and Windows) natively virtualized on top of our Xen-based virtualization platform to our early evaluation customers.

VI: In its last externalization, Red Hat accused Novell of not being responsible in delivering the product so early, compromising customers’ experience. What’s your take on this? Novell embedded the product too fast?

MG: Novell has always been about being first to market with their Linux distribution. Once again, they’ve pushed the envelope in their support for this specific capability and must feel their version of SLES is ready for their target customer. We applaud their efforts.

VI: Since the release of Virtual Iron 3 will base its own solution on Xen, what is the company’s position about Xen maturity? And mostly: why Xen is not ready? Where the problem lies?

MG: Xen, as an open source project, will never be enterprise-ready. That is the nature of open source. It’s the value that each vendor adds that makes the difference between a real product ready for production deployment and an open source project; between enterprise-ready and unstable.

Initially, some vendors thought that they could just take finished code from the Xen open source project and wrap it into their offerings. They were not prepared to make a significant development effort into the virtualization services layer or the virtual infrastructure management layer required to deliver an enterprise-ready product. These vendors dramatically underestimated what Xen was and what it wasn’t. In addition, they’ve come to appreciate just how complex virtualization software really is and the gaps between Xen the science project and a Xen-based solution that is ready for enterprise deployment.

As a focused virtualization solution provider, Virtual Iron's approach has been much more ambitious. We set out from the beginning to build an enterprise-ready virtualization solution on top of the Xen hypervisor and we brought a core competency in virtualization architecture and software development to the task. First we identified the gaps between what was in the project, and what we needed in it to deliver an enterprise-class solution. We worked closely with core members of the Xen development community – companies like AMD, HP, IBM, Intel, Novell and Unisys – and developed an architecture and robust code to support an enterprise-class solution. Virtual Iron extended the capabilities of Xen by building its own virtualization services and management layer on top of the Xen hypervisor. It’s now being evaluated at both end-user customer and strategic partner sites and is supporting enterprise-class workloads. We've attempted to contribute a lot of this code back to the open source Xen project because we want the user community to get these same benefits. This was a substantial development effort and it’s what really makes the difference between a Xen-based virtualization solution that’s enterprise-ready and those that are not.

Virtual Iron feels we have a version of Xen that is ready for evaluation as an Enterprise Data Center Virtualization platform. We think the difference is all in one’s market focus.

VI: Another point of interest lately has been the XenSource agreement with Microsoft for providing interoperability between XenEnterprise and upcoming Windows Server Virtualization. VMware severely criticized the move denouncing damage against open source community and the Xen project itself. What is the Virtual Iron position on this?

MG: It is great that Microsoft has embraced future paravirtualized Linux OS’s running on their future Windows hypervisor and hopefully the entire community will benefit from this expansion of the Xen project. It would be a huge red flag for the Xen project if this turns out to be a proprietary-only use of Xen.

However, many within the Xen community question the motives of XenSource:

• are they building a one-lane toll bridge between paravirtualized Xen Linux and the Windows hypervisor?
• is XenSource going to charge the Linux distributors to be able to run on the future Windows hypervisor platforms?
• is this similar in philosophy to XenSource’s recently announced Xen Certification of the community project?
• how do you certify a community project unless the real motivation is to leverage an open source community for commercial gains at the expense of the community? Only time will tell if this is a real open source project in the image of Linux or just another proprietary project cloaked in open source clothing.

VI: What does the future hold for VMware?

MG: We don't know exactly, but we can make some educated guesses based on the software industry's Darwinian history. For example, in the x86 server OS space, many different OSs have been squeezed out by Microsoft and the many different flavors of (Open Source) Linux. We expect a similar dynamic to occur in the server virtualization market over time. Longer term, it's not inconceivable that the VMware hypervisor is squeezed out of the market as this layer of the virtualization stack becomes increasingly commoditized and ubiquitous. Specifically, we believe that only two hypervisors will survive - proprietary (Microsoft) and open source (Xen-based). Only Microsoft is capable of sustaining the proprietary development efforts required. Open source represents the future of software development and is the only sustainable alternative to large-scale proprietary efforts.

From a business perspective, VMware has limited strategic choices as its ownership by EMC will put them increasingly at odds with their strategic partners. This is specifically why the Xen ecosystem emerged to begin with. In addition, Wall Street continues to question whether EMC will really invest in growing this business with complementary acquisitions, as it is not mainstream to their core business.

From a technology perspective, VMware's architecture is an eight year old, legacy way to virtualize the x86 architecture. Its approach will face continued challenges now that new hardware-assisted virtualization processors from AMD and Intel are shipping.

One week after the launch of the new esxMigrator tool, Scott Herold, Director of Research and Development at vizioncore, sit down with virtualization.info speaking about company products, future projects and state of the virtualization market.

virtualization.info: vizioncore just released a new product called esxMigrator, which helps customers to migrate from ESX Server 2.x environments to the new VMware Infrastructures 3 (VI3). Can you explain some of the issues raised by migrations and why esxMigrator is a desirable tool?

Scott Herold: The greatest challenge of migrating from ESX 2.X to VI3 is managing the downtime associated with each stage of the migration. VMware has made several significant changes to the underlying architecture of their virtualization platform that requires downtime of either your virtual machines or ESX hosts. Smaller environments have the issue of having to take a complete outage of their virtual infrastructure to perform an upgrade of their environment, which can take anywhere from 4-6 hours per host. Large VMware ESX implementations have the major issue of managing virtual machine and application dependencies to dictate their migration time frames.

To top it all off, VMware has changed their storage infrastructure to support a more organized subdirectory structure for VMDK files. Their current recommendation is to have all VMDK files for a single virtual machine reside on the same VMFS datastore in the same directory. That is the exact opposite of the ESX 2.X recommendation of balancing your workload across multiple LUNs and data paths. A lot of data needs to move to better organize it for the VI3 architecture. While it is being moved, it requires downtime.

esxMigrator automates and manages all of the steps involved in the migration and controls the transfer of data in the background while the source virtual machine is running. The final cutover of the virtual machine can be scheduled to occur at any point in time. This provides the maximum amount of control around a virtual infrastructure upgrade.

VI: Why would customers prefer to move to VI3 with esxMigrator instead of approaching a virtual to virtual (V2V) migration?

SH: In a way, esxMigrator IS a V2V migration tool. We take the unique approach of performing the V2V migration outside of the guest operating system. We require no agents inside or outside of a virtual machine and do not impact the resource utilization of the guest during the migration process. Using our methodology, a new virtual machine is built based off the ESX 2.X configuration and the tedious process of copying the data is performed on the backend leveraging technologies from our proven products. What you have at the end of the process is a fully VI3 compliant virtual machine all the way from the VMDK structure and layout to the VMware Tools and virtual hardware running inside the guest operating system.

VI: A new VI3 feature is VMware Consolidated Backup (VCB), able to backup running virtual machines acting at Storage Attached Network (SAN) device level. vizioncore has offered virtual machines live backup since its commercial launch with esxRanger. Would you clarify differences in esxRanger and VMware Consolidated Backup approaches?

SH: VCB, in its current state, is a backup platform. It enables ties to the vmkernel through a SAN to offload backup jobs from the console operating system. VMware is relying on backup vendors to supply the necessary ties to script the VCB technologies as pre and post activities of a standard backup job to perform the actual backup. There is no user interface surrounding the technology outside a set of command line executables.

While VMware does provide a mechanism to capture a full image of the server, they removed the necessary executables to easily restore the file from the GA release. There is also no easy way to restore individual files when using VCB in "File Mode". One of two things are required to restore data from a standard backup job: First, a backup agent can be loaded in each guest so they have the capability to restore individual files. This defeats the entire purpose of VCB. Second, a proxy server can be used to manage the restoration of files. The data may then be copied back to their original locations on their original virtual machine. This adds a layer of complexity to the backup/restore procedures of an organization. A final major pitfall of VCB and how it is structured is that file permissions are stripped from any file or directory that is backed up through the proxy. This is a major issue for any organization concerned with securing their data.

esxRanger will have ties to VCB to better manage the use of the platform through an intuitive user interface and enhanced data management, but will still have the capability to perform image level backups across the entire platform. iSCSI, NFS, and local VMFS targets will not be an issue for esxRanger with VI3 compatibility. With the recent addition of file level restore capabilities, esxRanger will provide a complete package for restoring both images and individual files to any storage platform.

VI: In previous months vizioncore launched a much appreciated free package called esxBasics, bundling a scaled down version of esxRanger and esxCharter. Is there any plan to extend the bundle to esxReplicator and esxMigrator?

SH: As of right now there are no immediate plans to include our newer products into the esxBasics package. The way esxReplicator and esxMigrator are licensed would provide a challenge for us to properly manage a "Basics" version. With the introductory pricing of esxMigrator that we have announced, we are not anticipating this to be a significant issue. We will continue to retrofit beneficial technologies into esxBasics as we see fit, but will always ensure our esxRanger processional and standard customers have all the functionality they require first.

VI: A part of VMware customers adopting GSX Server have no opportunity to address virtualization issues with vizioncore products because you are currently offering solutions for ESX Server only. Now that GSX Server became Server and is offered for free that customer base could grow a lot: does vizioncore plan to introduce products for the new platform? If so, only on Linux or also on Windows hosting?

SH: vizioncore is actively looking at opportunities in the VMware Server space. We anticipate a fast and wide adoption of VMware's free offering and are analyzing how our products and ideas would best fit this technology. From a development standpoint, there is almost no difference between managing a Linux or Windows hosted VMware Server platform. There are minor differences around file system structure between the platforms, but those aren't even speed bumps for the developers we have on staff.

VI: Is there a chance vizioncore will broaden its product line to different virtualization plaforms like XenEnteprise or Virtual Iron?

SH: vizioncore has a strong partnership with VMware and will maintain a strong focus on enhancing VMware's technologies. That does not mean we aren't very well aware of the other technologies in the virtualization space and what opportunities may be available for each of these platforms. It would not be a wise business decision on our part to lose focus on all available technologies as virtualization platforms become a commodity.

VI: vizioncore is expected to release a new version of esxRanger within middle August, introducing support for VI3 and file level restore. What will we see next? Can you provide a roadmap of features vizioncore is to deliver?

SH: We are working on several items in the office. One of our main areas of focus is on enhancing VCB. There are a lot of opportunities to further strengthen VMware's offering. We will also be implementing a differential backup engine in the near future. We have developed a unique process that can be applied to multiple operating system platforms and file types. We will still be able to leverage individual file level restores out of our differential backups, which is a very unique feature. We are also members of VMware's Community Source program and have been working closely with VMware on some extremely exciting technology that is going to change the way backups and replication are currently performed.

VI: Why do you think VMware is so slow providing out of the box disaster recovery features like platform redundancy and virtual machines live backup, even if the virtual infrastructure greatly aggravate the concept of single point of failure?

SH: VMware has a strong ISV program and has a lot of support from their partners in assisting with developing this technology. We see VMware making the right steps in some of these directions in VI3 with offerings such as VMware HA, DRS and VCB. The technology involved with developing around the VI3 architecture is extremely advanced. VMware is a very rapidly growing company and may not want to spread themselves too thin in various projects. There is also a lot of new intellectual property surrounding a lot of the methodologies of advanced disaster recovery and availability. Legally, it's a mess tracking all of the new ideas and processes. It also tends to take time to buy technologies or certain aspects of it to integrate into a platform as advanced as VI3.

VI: In your opinion which is the feature customers are missing the most in today's virtual infrastructures? Is vizioncore working on a product about it?

SH: When I was in consulting, the biggest problem I noticed was controlling the growth of a virtual infrastructure. The time to deploy a new operating system sharply fell from weeks to hours. As soon as a company sees the true benefit to virtualization, the flood gates open. Organizations no longer have "Server Sprawl", but they still have "OS Sprawl". Introducing a large number of operating systems, whether physical or virtual, provides significant challenges around all aspects of IT management. Security, Networking, Storage, Support... all of it becomes more difficult as more and more systems are added to the environment.

If I had to pick the single component that I feel needs the most consideration, it would have to be the security aspect around virtual machine management and deployment. While there is nothing on our immediate roadmap we are making progress through VMware's Community Source program to enable us to do some amazing things with data. That will open a lot of possibilities for us around ways to manipulate data that have yet to be identified as we progress through community source development.

VI: vizioncore appears to have a very fast development lifecycle. How much part of research is driven by customers needs? How much time do you spend, as Director of Research and Development, in the virtualization community?

SH: I spend more than 50% of my time in the community. One of the things that I have found has changed is that I have moved from being a very active member of the public community to a very active member of the private community. A lot of the things I do now are invisible to most people who watch community news or VMTN forums, etc. I have implemented a customer visit program for my department where we head out to our larger customers to see how they are using the product and make sure they are following not only our own, but VMware's best practices in their infrastructure. The amount of product feedback we can receive from even a single organization is pretty significant.

I still participate in user groups and various conferences such as VMworld, and love to get out and just give people real technical information without the marketing. On the private side I keep my eye on every aspect of technology whether it currently touches virtualization or has the potential to in the future. With the current growth trends in virtualization, the best is still yet to come. vizioncore has a very large head start over others competing in the same space as us.

Two weeks ago Microsoft and XenSource announced an agreement to grant interoperability of virtual machines on upcoming Windows Server Virtualization and XenEnterprise virtualization platforms.
The move raised the attention of the whole IT world, involving licensing, supporting, security and performance issues.

virtualization.info interviewed both companies to further understand details of the agreement and spread some lights on what customers have to expect for the Microsoft hypervisor release. To answer questions I met Mike Neil, Senior Director of Virtualization Strategy, Windows Server Division, at Microsoft, and Simon Crosby, CTO at XenSource.

To simplify questions and answers since now we'll call a virtual machine natively running on Microsoft hypervisor, Windows Server Virtualization, as WSV-VM and a Xen virtual machine natively running on XenSource hypervisor, XenEnteprise, as XE-VM.

Microsoft Side

virtualization.info: The WSV-VM running within XenEnteprise will have all capabilities it already has on Windows Server Virtualization or there will be some limitations?

Mike Neil: The technology resulting from this agreement will provide interoperability between Xen-enabled Linux guest operating systems running on Windows Server virtualization in Windows Server Longhorn. Windows guest running on XenEnterprise will continue to function in the same way they do today. For customers with Premier-level support agreements, Microsoft will use commercially reasonable efforts to investigate potential issues with Microsoft software running in XenEnterprise or other non-Microsoft virtualization technology. Our product support policy is described here.

VI: The upcoming Virtual Machine Manager will be able to centrally manage XE-VMs along with WSV-VMs, including tasks like provisioning and live migration between hosts?

MN: The first release of System Center Virtual Machine Manager is focused on the management of Windows environments. That said, it will be able to centrally manage Linux guests.
Some of the things you can do with SCVMM and Linux guests are:

• Deploy Linux VMs in VHD format from the central library
• Configure the virtual machine parameters for Linux VMs, such as RAM and disk space
• Control the state of the virtual machine (start/stop, pause/resume, save/restore)
• Live migrate a running Linux VM from one physical host to another

These are features that are enabled by the System Center Virtual Machine Manager and Virtual Server 2005, or Windows Server Virtualization, and are not specific to the technology being developed as part of the XenSource agreement. Live migration of guests will be a feature for Windows Server Virtualization.

VI: Will Microsoft offer support for XE-VMs running on WSV? If so will it be equal to one offered for Linux guests natively created on WSV?

MN: We currently support Linux running as a guest in Microsoft Virtual Server 2005 R2 from both a technology perspective and a 24-hour technical support perspective. When we added the support of specific Linux distributions on Virtual Server 2005 R2, we made a long-term commitment to make sure that non-Windows operating systems can be run in a supported manner, both on top of Virtual Server and our future virtualization products. Proving support for XE-VMs on Windows Server Virtualization is a part of the commitment.

VI: In a scenario where a WSV-VM is moved on XenEnteprise hypervisor some critical issues raise about licensing and support. If the Microsoft customer is using an unlicensed version of Windows inside the WSV-VM (which is permitted by the new virtualization licensing model up to 4 virtual machines), what will happen when he'll run it on XenEnterprise? And who between Microsoft and XenSource will support the virtual machine on that case?

MN: The four virtual instances allowed with Windows Server 2003 Enterprise Edition are not unlicensed. They are in fact license rights granted for the Enterprise Edition. We also recently extended virtualization licensing rights to Windows Server 2003 Datacenter Edition, providing for unlimited virtual instances with that version of Windows Server.

A customer running Windows Server needs to acquire a license for the physical machine they are running Windows on, regardless of the virtualization technology they are using. If they are running a VM on XenEnterprise, they need to have acquired a license for that Windows machine as well. More details about Microsoft's virtualization licensing policy can be found here.

As I mentioned above, for customers with Premier-level support agreements, Microsoft will use commercially reasonable efforts to investigate potential issues with Microsoft software running in XenEnterprise or other non-Microsoft virtualization technology.

VI: How much time the agreement will last?

MN: The agreement is to develop and deliver the technology for interoperability between Xen-enabled Linux guests and Windows Server virtualization, to be delivered around the same time Windows Server virtualization is delivered (within 180 days of Windows Server Longhorn, which is slated for release in H2 2007). That said, we do have an ongoing relationship with XenSource (they licensed Microsoft's Virtual Hard Disk format, for example). We share a common goal to help customers more easily realize the benefits of virtualization.

XenSource Side

virtualization.info: Technically speaking what will happen exactly in the interoperable scenario, where a XE-VM will be executed by Microsoft Windows Server Virtualization hypervisor?

Simon Crosby: Our announcement of a strategic partnership with Microsoft will enable Xen-enabled Linux guests to run with full benefits of paravirtualization (Microsoft terms it enlightenment) on the upcoming Windows Hypervisor, code named Viridian. Viridian and Xen share a common architecture, and are both paravirtualizing hypervisors. This is key, because paravirtualization has been recognized as the most important enabler of virtualization by every OS vendor. This architecture will be supported in the next release of every x86 OS of relevance to the enterprise, with RHEL 5 and SLES 10 Linux incorporating Xen (indeed many other Linux distributions will do so too); Sun with a Solaris 10 update expected later this year, and Microsoft with Viridian. Viridian will ship as an embedded component of Windows Server Longhorn.

When the Xen-enabled Linux guest runs on Viridian, it will use the native Xen hypercalls to access virtualization functions. A small adapter will adapt the Xen hypercalls into Viridian hypercalls so that the Linux guest can run with full performance on this paravirtualizing hypervisor.
In addition, we are collaborating with Microsoft on the delivery of paravirtualizing I/O capabilities for Xen-enabled Linux guests to run on Viridian. These capabilities are called Virtualization Service Clients and Virtualization Service Providers in the Microsoft terminology, and they correspond to the Xen front end and back end drivers used for paravirtualized I/O.

VI: The XS-VM running within Microsoft WSV will have all capabilities it already has on XenEnteprise or there will be some limitations?

SC: To be clear: The Xen-enabled Linux will be exactly the same Linux as is shipped by Red Hat or SUSE, or whatever other distribution we support. This has nothing in particular to do with Xen Enterprise, which is XenSource's product and which also supports those Linux guests. That is, all implementations of Xen will support these Linux guests, since all Xen implementations (including in Solaris, RHEL, SLES) support paravirtualized guests.

The answer to the question is thus: The Xen enabled Linux guest will have all of the capabilities that it has when running virtualized on Xen.

VI: In this collaboration will XenSource have full access to all Microsoft WSV software code? If not, how XenSource will be able to assure that performance and security levels of a XE-VM will be identical on both XenEnterprise and Windows Server Virtualization?

SC: The terms of the collaboration between XenSource and Microsoft have not been disclosed, however we certainly can state that we have a license from Microsoft to implement the adapter against the Viridian hypercall API. Microsoft has disclosed that API to several vendors, and discussed it in detail at the recent WinHEC conference.

VI: Is this agreement breaking in some way GPL license of Xen or Linux in general?

SC: Not at all. XenSource is committed to the GPL Xen code base as the key to our powerful community and the ubiquitous delivery of a uniform feature set. We lead the industry through open development, and every feature that is currently in development, that does not require a closed source license by virtue of a license agreement with a 3rd party, is targeted for GPL implementation.
It is XenSource's intention, wherever possible, to deliver features into the open source Xen code base. In the specific case of our Microsoft partnership, some components cannot be released under GPL.

VI: Can we expect some support available to the open source community, from know-how XenSource will gain during this agreement?

SC: The benefits for Xen from our collaboration with Microsoft will be tremendous. We anticipate that we will be in a far better position to deliver high performance Windows support on Xen, and moreover Microsoft has agreed to support Windows on our own product, XenEnterprise.
More importantly, perhaps, is the recognition from Microsoft that Xen's paravirtualization (enlightenment) is the hypervisor architecture of the future, as opposed to emulation and binary patching. It is a vindication of the technology leadership of the Xen community, of our open source collaborative development and ubiquitous deployment. It is an acknowledgement from the most powerful OS vendor in the industry, that the Xen hypervisor is the hypervisor to beat, and that the installed base is about to be consigned to history. A common architecture, supported by every OS vendor in the industry, is emerging.

Xen's paravirtualization architecture has been endorsed by every OS Vendor, and this is a tremendous shot in the arm for every vendor in our ecosystem. Microsoft's support of that will be of great importance to every vendor working on Xen - as it is a recognition that their investment in an open industry standard platform has been proven worthwhile.

Immediately after the release of new Virtual Infrastructure 3, virtualization.info had the pleasure to reach Raghu Raghuram, Vice President of Platform Products at VMware, to ask him details about just launched products, recent Akimbi acquisition announcement, secret VMware Integrity product, planned presence in the Apple operating system and further steps in virtualization market leader strategy. From his answers a revelation comes out: VMware is going to partially support Microsoft virtualization technologies.

virtualization.info: The VI3 beta program has been highly successful, even involving enterprise class products, often with complex testing infrastructures to setup and maintain. Why do you think the beta program received so much participation?

Raghu Raghuram: Two reasons:

• VI3 breaks new ground in terms of capability and delivers full infrastructure virtualization. Customers were very eager to experience the new functionality and understand how they can benefit from it.
• Virtualization has become mission-critical. More than 90% of our customers run in production, 2/3rds are relying upon our products for disaster recovery and over half our customers use VMotion everyday to minimize planned downtime or respond to load fluctuations in the data center. So these customers want to be sure that the new software works well in their environment and is robust enough for their production needs. We were very gratified that an overwhelming percentage of the beta customers declared the product to be robust and high quality before we shipped the final bits.

VI: Looking at VMTN Forums the most visited thread about VI3 is called "Comparing ESX 2.5 to 3.0 performance". On it some customers are reporting a performance increase of up to 50%. Can you confirm such result? Are they achievable only with special hardware configurations?

RR: We have done a huge amount of work in increasing the performance and scalability attributes across the board. It is great to see that customers are experiencing the improvements. An upcoming blog from Steve Herrod, our VP of R&D describes some of these changes and improvements. As for specific numbers, they depend upon the type of workload and hardware configurations. We are committed to being the leader in this space not only in terms of functionality and robustness, but also scale and performance. So you will see continuing emphasis on these areas.

VI: A lot of confusion has been generated around the Virtual Machine File System (VMFS) option in new licensing model. Customers buying the VI3 Starter Edition believe they are loosing the famous VMware file system. Would you clarify this point?

RR: VMFS provides two distinct set of benefits. First, as a Virtual Machine Filesystem, it is optimized for storing VM disks as files. This capability is available across all product editions. Second, VMFS also serves as a distributed file system that connects to Fiber-channel SAN storage. VMFS, in this mode, offers capabilities for volume management, handling heterogeneous storage arrays, capacity expansion, on-disk locking to enable concurrent access by multiple ESX Servers. This aspect of VMFS, or clustered VMFS as we call it, is critical enabler of advanced functionality like VMotion, DRS with SANs. Clustered VMFS is available with VMware Infrastructure Standard and Enterprise.

VI: A significant objection to VI3 release relates storage: in several press articles customers' interviews underlined how VMware strategy is still unable to address a compelling need of flexibility in storage management. What do you think about this point of view?

RR: VI3 actually does advance the capabilities for storage management through the support of SMI-S APIs in ESX Server 3.0. Our strategy is to leverage and drive open industry standards for management of virtualized infrastructures and we will continue to do that for storage space as well.

VI: The new VI3 brings in powerful capabilities but also more complexity. Is VMware going to renew its certification program and training offering? When we can expect new courses?

RR: Absolutely. You will find the details at http://www.vmware.com/certification

VI: The biggest news these days apart the VI3 launch and availability is the Akimbi acquisition. Slingshot technology will be released as stand-alone product, available for both VMware ESX Server and Server, or as a module for VirtualCenter?

RR: Akimbi technology will be incorporated into a standalone software lifecycle management product, available for both VMware Infrastructure and VMware Server.

VI: Many Akimbi customers are now asking what will happen to the Microsoft part of the product. VMware is going to shutdown it as expected? Or VirtualCenter is going to expand its capability managing other virtualization platforms?

RR: VMware will support Microsoft Virtual Server 2005 in its software lifecycle management product.

VI: Another hot news involves something virtualization.info has discovered last week, dubbed VMware Integrity. Can you tell us something about this one?

RR: We've been proactively evangelizing the value of virtualization in providing better solutions for security - it happens to be one of the primary areas of research of our Chief Scientist Mendel Rosenblum. We've got many interesting technologies in our labs that we believe will change the way people manage security just like VMotion changed the way people think about server availability, but at this point we don't have anything more specific to comment on.

VI: A very interesting product VMware is not updating since a while is ACE. Can customers expect something new within this year or the product is going to be dismissed?

RR: We are excited by the potential of ACE to improve desktop manageability and security. We are hard at work on the next major version. As the product is readied for beta we will offer more details.

VI: Parallels Inc., the last company entered in the virtualization market, just released a desktop virtualization product for new Apple Mac OS X for Intel architectures, beating any competitor on time. Customers are a bit surprised the market leader, founding its success on a high quality desktop product like VMware Workstation, is still mum about a possible presence in the Apple operating system. Isn't VMware interested in Apple market?

RR: With Apple switching to x86-based processors, robust and proven virtualization capabilities for Apple users is an interesting opportunity. We have stated that we do have VMware running on Mac OS X in our labs - stay tuned for future announcements in this area.

VI: A last question about Microsoft competition. Where VMware will be at the time Microsoft will release Windows Server Virtualization and Virtual Machine Manager, 2 years from now?

RR: The first generation of virtualization was single-server partitioning. The second generation of virtualization provided a layer of management for Virtual Machines. With VI3, we have entered the third generation of virtualization. Virtualization has evolved from being a static single-server partitioning technology to a distributed, infrastructure-wide virtualization technology that provides the foundation for next generation systems infrastructure. The next generation systems infrastructure will harness a collection of industry-standard hardware components and deliver the utilization, reliability, availability and security that was previously possible only with million-dollar mainframes. VMware's mission is to enable this next generation of systems infrastructure even as other vendors focus on basic partitioning and management two years from now.

The last week announcement of new Microsoft virtualization strategy at WinHEC 2006 conference brought in big interest.

The importance of announcement and the technical complexity of products (the so-called Windows hypervisor will be integrated in the operating system itself) also raised customers confusion, now trying to understand how new products will address problems today they are solving with Virtual Server.

Mike Neil, Virtual Machine Technologies Product Unit Manager at Microsoft, accepted to sit down with virtualization info and clarify some important points about Windows Server Virtualization (WSV), Virtual Machine Manager (VMM), Softricity acquisition and more, in an exclusive Q&A session:

virtualization.info: Windows Server virtualization (WSV) is going to be delivered as stand alone package, within a Service Pack, or will be part of a possible Longhorn R2 version?

Mike Neil: The packaging and ship vehicle for Windows Server virtualization is still to be determined. Windows Server virtualization is being developed as an integrated part of Windows Server Longhorn and will be delivered within 180 days of Longhorn RTM. We'll update customers and partners prior to the beta release.

VI: Windows Server version will be able to offer WSV or customers will need a special version? Something likes Windows Server 2007 Virtualization Edition?

MN: Windows Server virtualization is a platform technology that will be an integrated component of Windows Server "Longhorn" so it'll be available as part of the operating system.

VI: Windows itself will be installed inside a WSV virtual machine? If so there will be a way to avoid using it?

MN: Windows Server virtualization is composed of three major architectural components: the Windows hypervisor, the virtualization stack and the new virtualized I/O model.

The Windows hypervisor runs under all of the partitions and provides the very basic virtualization services to create secure and isolated partitions. The virtualization stack runs as a component of Windows in the parent partition, and the virtualized I/O model has components that run the parent partition and communicate with counterparts in the child partition to provide enhanced I/O capabilities.

For a user to be able to create and run virtual machine and get all of the features and value of Windows Server virtualization they will use all three integrated together.

VI: The new Server Core feature of Longhorn will be able to offer a WSV role?

MN: Good question and one that we're working on to finalize so we'll have to get back to you later. There are certain roles available in Server Core right now and we are evaluating what other roles might be available in Server Core installations for RTM.

VI: WSV will be able to import virtual machines made with other virtualization platforms (VMware products and Xen in particular)?

MN: We are not planning to convert other virtualization solutions directly, but we are enabling third parties to provide customer solutions.

Microsoft is licensing our VHD format royalty free and will fully document the new WMI interface that we are developing as an industry standard through the DMTF. With this information, third parties can provide solutions that can convert other virtualization formats into a format that is compatible with Windows Server virtualization.

Additionally, users of Microsoft Virtual Server 2005 will be able to migrate to Windows Server virtualization, preserving their investment in Virtual Server today.

VI: WSV will feature high availability features, like Virtual Server 2005 already do today? (I'm talking about host OS failover capability)

MN: Windows Server virtualization will offer many high availability features to help customers produce a resilient platform for consolidation and business continuance.

Windows Server virtualization will offer guest-to-guest clustering for cluster aware applications. It will also offer host-to-host clustering that allows the automatic failover of all virtual machines to a backup host should the primary host fail. And it will allow the live migration of a running virtual machine from one host to another with little or no downtime, which can be used for servicing a host for planned downtime.

Windows Server virtualization will leverage the significantly enhanced clustering capabilities of Longhorn Server. These include a new best-of-both-worlds quorum model that produces a hybrid of Majority Node Set (MNS) logic and Shared Disk Quorum model. This new hybrid design allows geographically dispersed clusters with or without shared disks.

VI: Is WSV adhering virtualization standards VMware proposed and other vendors are considering? Or Microsoft is going to launch its own standardization?

MN: Microsoft is working with industry organizations and partners to foster interoperability and standards on many fronts:

• Microsoft is working with the industry in the DMTF on standards for virtualization management. As part of this effort we are working with HP, IBM, XenSource and VMware as well as other partners in the industry
• Microsoft is working on industry hardware standards for device virtualization through the PCI-SIG
• Microsoft also provides royalty free licensing of our virtual hard disk (VHD) format
• Microsoft will license the hypercall interface for the Windows hypervisor and a preliminary version of that interface will be provided to WinHEC attendees on their conference DVD

VI: Will WSV be able to provide hot modification of virtual hardware also for non-Windows guest OSes?

MN: Windows Server virtualization allows for the hot-addition of RAM and CPU resources to a running virtual machine. This is done through industry standard mechanisms as part of the ACPI interface. It is then dependant on the non-Windows guest operating system to supports these ACPI interfaces correctly.

VI: Has Microsoft any plans to support other operating systems (Sun Solaris 10 in particular) with WSV?

MN: We have no plans to speak of today. We'll continue to poll our customers and partners for input on interoperability and expanding guest OS support within Virtual Server and Windows Server virtualization.

Microsoft is making our hypercall interface available to other operating system vendors if they wish to enhance their operating system to take advantage of the Windows hypervisor. Conference attendees of WinHEC will receive an draft of the hypercall API on their conference DVD.

Mike, now let's try to clarify some points about the announced System Center Virtual Machine Manager (VMM), previously knonw as codename Carmine.

VI: VMM will be mandatory for WSV administration? If not, what customers will be able to do without it?

MN: Windows Server virtualization will come with a new MMC-based UI that was demoed during the Bill Gates keynote at WinHEC.

This new UI allows the administration of virtual machines on a host running Windows Server virtualization either locally or remotely. System Center Virtual Machine Manager will provide data center level virtual machine management features integrated with the System Center suite of management products. System Center Virtual Machine Manager will mange both Microsoft Virtual Server and Windows Server virtualization hosts.

Additionally, Microsoft is working with a broad set of management partners to ensure that Windows Server virtualization and Longhorn Server can be managed.

VI: There are any plans to provide a VMM version not requiring Active Directory?

MN: System Center Virtual Machine Manager beta 1 will require Active Directory and the initial thinking is for the final released product to be the same. However, the primary reason for getting the software into customers' hands in beta form is to identify opportunities to improve the software.

Active Directory is currently required for security purposes to ensure that access to images on the network is authenticated and secure. Active Directory also provides benefits by allowing the SC VMM administrator to identify and understand all the physical servers running virtual machines that are installed in their environment and the use of Distributed File System to replicate images automatically across the network.

VI: Which opportunity will have partners to add value to VMM?

MN: We are working closely with our partners to help them understand where we are making technology investments and how they can help extend and enhance the solutions we deliver in this space….such as V-to-V and V-to-P migration, non-Windows OS configuration and image transport over SAN infrastructure. Since System Center Virtual Machine Manager is based on top of Windows PowerShell - the new command line shell and scripting language for Windows - there are many opportunities to enhance and orchestrate its functionality through Windows PowerShell.

Mike now I'd like to ask you something about another couple of hot topics: Softricity acquisition and offering for Apple Mac OS.

VI: Is the Microsoft acquisition of Softricity confirmed?

MN: Our intent to acquire Softricity was announced at WinHEC. The acquisition is expected to close 30-45 days.

VI: What's the strategy about application virtualization? I mean: do you plan to deliver a unified platform offering application and server virtualization at the same time?

MN: Our goal is to provide an integrated, comprehensive and cost-effective virtualization solution for our customers.

We are making significant investments both at the platform and management level to accomplish this goal. We look at virtualization at three different levels - machine, OS and application -- as each level provides different levels of isolation and granularity.

Machine virtualization provides the most isolation with the least granularity, while application isolation provides the least isolation with the most granularity.

Softricity provides application virtualization that complements our other virtualization technologies - Virtual Server, Virtual PC, Windows Server virtualization and System Center Virtual Machine Manager.

With this addition, Microsoft is able to provide customers with a broad range of virtualization technologies to choose from, to achieve the right balance between flexibility, isolation and performance.

VI: If so is this going to happen for the launch of Windows Server Virtualization or there will be a further delay to achieve the task?

MN: The Softricity acquisition doesn't impact the schedule of Windows Server virtualization.

VI: You probably know a company called Stream Theory just filed a lawsuit against Softricity and others application virtualization companies for application streaming technology. Is this going to impact in the acquisition process?

MN: I'm aware of this matter, but it's not appropriate to make any assumptions or statements with respect to it.

VI: Few months ago Microsoft declared its intention to offer a virtualization solution for the new Apple Mac OS X on Intel architecture. Are you still on this path? If so there are raw dates for release?

MN: We haven't declared intentions to develop a virtualization solution for Intel-based Macs. Microsoft's Mac business unit remains in the best position to provide a fully integrated non-dual boot solution that works with Windows. This team is working with Apple to incorporate changes in the OS that could allow a Virtual PC solution to work on Intel-based Macs. We're still in the process of investigating a universal version of Virtual PC.

Mike, the last one question is about another virtualization technology Microsoft seems to be interested in: OS partitioning.

VI: Can you tell virtualization.info something more about Microsoft efforts in OS partitioning you mentioned one month ago?

MN: During his WinHEC keynote, Bob Muglia stated that Microsoft is making some long-term investments in this space.

It will be a few years before you see these investments come to market, but we think that OS level virtualization is a very important component in our overall virtualization strategy. It provides an additional option for customers trying to balance isolation and scalability.

virtualization.info starts 2006 with an exclusive interview to Benjamin Rudolph, Marketing Manager at Parallels.

Parallels officially released its Workstation 2.0 product at the end of 2005, entering in the desktop virtualization market where VMware, Microsoft and Serenity Systems International are already.

In the following interview Benjamin Rudolph talks about Parallels 2006 roadmap, mentioning enterprise virtualization products and touching hot topics like Microsoft Vista Aero support and Apple MacOS x86 virtualization.

virtualization.info: In a previous post virtualization.info reported that Parallels was eventually involved in the development of a virtualization project launched at the beginning of 2004, known as TwoOSTwo and then renamed SVISTA, owned by a company called Serenity Systems International. Would you spend some words to describe the company history and explain this mystery?

Benjamin Rudolph: Parallels has its own, proprietary code base and an independently developed product line. We have no current relationship with Serenity.

VI: You just hit the market with a new product: Parallels Workstation 2.0. It seems to have two interesting features to compete: a very cheap pricing and the technology primacy of embedding a hypervisor in a desktop product. How the hypervisor distinguish Parallels product from competitors?

BR: During Parallels Workstation's development cycle, we took the hypervisor concept from server virtualization and made it work on the desktop, and the result was the world's first hypervisor-powered desktop virtualization solution. Parallels Workstation's lightweight hypervisor inserts a thin layer of software between the host computer's hardware resources and the primary OS that directly controls some of the host machine's hardware profiles. This technology significantly improves virtual machine stability, performance and isolation by enabling virtual machines to directly access hardware resources, rather than having to first pass through the primary operating system.

Also, Parallels' lightweight hypervisor also enables Parallels Workstation to be the first desktop virtualization solution that fully supports all of the features and benefits provided by Intel Virtualization Technology ("VT"). Intel VT is a "built-in" hardware virtualization technology that optimizes a CPU for virtualization. Intel VT will soon be a standard component of Intel's CPU offerings, which indicates just how important virtualization solutions, like Parallels products, are to the future of computing. When using Parallels Workstation on VT powered system versus a standard system, users will find that virtual machine performance can increase several times! AMD is also developing its own hardware virtualization technology, which is code-named "Pacifica." Parallels products will also offer full support for this technology when it becomes available to the general public.

I think it's important to note that Parallels Workstation's hypervisor is just one of many important features that users will find very useful. The product also offers one-click installation and setup, an intuitive web-like control panel that anyone can easily use, strong primary and guest OS support that includes Windows, Linux, FreeBSD, OS/2, eComStation, and MS-DOS, a compact, resource-friendly download and installation package, and an industry leading price of $49 per license. All in all, the product is the easiest to use, most cost effective, high-performing solution available today.

VI: A great missing of today virtualization products is offering host OS support for Sun Solaris 10 operating system. Is Parallels Server going to provide such a support from the beginning?

BR: We are currently discussing the possibility of including Solaris 10 as a primary OS. We would definitely be interested in talking to Sun about building Parallels products that work seamlessly with this important operating system.

Parallels Server will definitely offer support for Windows and Linux primary OSes, as well as all of the guest operating systems that are currently supported by Workstation (Windows 3.1 - XP/2003, Linux, FreeBSD, OS/2, eComStation, DOS). Parallels Server will also offer guest OS support for Novell Netware, MacOS x and of course, Sun Solaris.

VI: The next big step about guest OS virtualization could be providing support for the expected MacOS for x86 architecture. You just said Parallels plans future support for it: are you talking with Apple to achieve this big mission?

BR: We are planning to support MacOS X as a primary and guest operating system in upcoming versions of Parallels Workstation. Unfortunately, I'm unable to discuss any technical or business relationships relating to this support.

VI: Apart MacOS for x86 another very expected OS for virtualization is the upcoming Microsoft Vista. As you know the new GUI environment, Aero, will require a lot of video RAM, which virtualized video cards actually don't have. So a Vista guest OS won't show you all new graphical enhancements everybody's waiting. Are you planning something to solve this?

BR: We plan to offer support for Microsoft Vista as a primary and guest OS in the very near future and support for the new "Aero" GUI will not be a problem. Parallels virtualizes generic VESA 3.0 video card in which video memory isn't fixed in a guest card specifications. Rather, it is allocated dynamically based on the demand of the appropriate guest video mode. Because of this capability, Parallels' VESA 3.0 emulation can be easily adjusted to suppport "Aero".

VI: Talking about competitors: Parallels just entered in one of the most hot IT market of the coming years and need to immediately face two huge animals. One a side you have a fast and furious gorillas, VMware, who is on this segment since ever and is actually the most innovative and recognized virtualization company around. On the other side you have a huge elephant, Microsoft, who is moving very slowly but with a powerful potential impact. As if this wasn't enough a third sharp tiger, Xen, is going to offer full and high performing virtualization at a small cost even on the Microsoft side. How Parallels plan to emerge on this scenario?

BR: Compared to other available virtualization products, Parallels products are much easier to use. Our one-click installation and intuitive, web-like interface enables anyone to use the product, even if they're not a "computer pro." Users won't have to struggle through a multi-step installation process or a complicated setup. With support for Windows, Linux, FreeBSD, OS/2, eComStation, and DOS (and future support for MacOS X, Solaris, Netware, QEMU, Plan 9, and many others), our primay and guest OS support is among the strongest in the industry.

It's also important to recognize the benefits of purchasing an affordable Parallels solution compared to an open-source product. Because all of our server and desktop virtualization products are developed entirely in-house using proprietary code and technologies, we are able to offer users caring, responsive support and friendly customer service in addition to great products.

On the business front, Parallels is aggressively expanding its channel sales program. By partnering with leading distributors and resellers around the world, Parallels will be able to extend its reach to all levels of users. We are also actively pursuing OEM and ISV partnerships. The goal of these partnerships is to make Parallels virtualization a "default" technology that is a critical component of a computer users daily operations, like a web browser or word processor.

VI: On the Workstation 2.0 press release Parallels announced server products launch for 2006. virtualization.info wants to know as much as possible about this. What kind of products we should expect? In which timeframe?

BR: As you know, Parallels Workstation is the company's first commercially available product, which reached the market in December 2005 after the conclusion of a highly successful beta program that involved thousands of users. So far, response to the release of the product has been outstanding. More than 100,000 users have downloaded the product since its launch on December 8th, 2005!

In Q1 06 we plan to release Parallels Workststation 2.1, followed by version 3.0 of Parallels Workstation in mid 2006. While the new feature sets in development for these releases are confidential, I can tell you that we're working hard to address the comments, concerns and "wish lists" that we've received from our beta and registered users. I'll be happy to discuss 2.1 and 3.0 features with you when we're ready to launch their respective beta programs. Beta testing for 2.1 should begin in late January, so we'll be able to share details soon!

In mid 2006, Parallels will be releasing its first server virtualization solution, Parallels Server. Parallels Server is an efficient server virtualization solution specifically designed to address the needs of small and medium businesses. The product enables users to create multiple self-contained virtual servers on one physical server. Parallels Server-powered virtual servers can efficiently run nearly any x86 operating system.

Due in late-2006 is Parallels Enterprise Server, a high-power server virtualization and management solution that installs on bare hardware, pools hardware resources and dynamically allocates them to virtual servers as necessary. This approach to server virtualization ensures that each physical server is used to its maximum potential, and that each virtual server always has the resources it needs to operate efficiently.

As you can see, we have a very aggressive product roadmap that compliments our mission of bringing virtualization to everyone. It's going to be a great year!

VI: Parallels Enterprise Server seems very interesting. Are you going to offer it just on certified hardware or will it be developed to be installed on generic x86 hardware?

BR: Parallels Enterprise Server's unique architecture will enable it to install directly on bare metal, as well as work with nearly any server hardware configuration.

VI: What about "secondary" virtualization features like Physical to Virtual (P2V) migration, other virtualization products VMs conversion or centralized hosts and guests management? Is Parallels going to release any products in these areas in 2006? Or should we expect a SDK to permit 3rd parties to develop them?

BR: Parallels has an exciting roadmap for 2006. As I mentioned earlier, our in-development feature sets are confidential until we launch each product's beta program. All of the tools that you mentioned are important features for our products to have, and we are working on a number of partnered and in-house solutions that will address them in the best way possible.

VI: Virtualization customers usually want three core features: reliability, performances and strong support. Why Parallels technologies should be the choice of customers looking for these features?

BR: Parallels products' lightweight hypervisor, small program footprint and full support of Intel VT technology offer users strong virtual machine stability and high performance, while our broad hardware and OS support ensure that any user can embrace virtualization, regardless of their hardware or software configurations. All of these great features, as well as the product's unmatched ease of use, are all packaged together with an industry-leading price point. We're 100% committed to making sure that using great software doesn't break your budget or empty your wallet!

Its also important to note that our products offerings are going to continue to improve. Our engineering team is working tirelessly to expand our OS and hardware support, improve virtual machine speed and performance, and make each product's user interface as easy to use as possible.

Look for big things from Parallels in 2006!

Steven Pollack, CEO, Platespin, accorded a technical interview with us about PowerP2V features. A lot of news emerge about this, exclusively for virtualization.info readers:

virtualization.info: Physical to virtual (P2V) migration is a probably the most needed process for obsolete datacenters IT managers embracing virtualization technology. But available tools on the market right now seem highly expensive. Would you give us some raw times? How much time is needed for a PowerP2V conversion and for a manual conversion of a legacy Windows NT 4.0 box?

Stephen Pollack: We find many customers look for cost justification based on the structure of their ITorganization. For example, several customers look to PlateSpin PowerP2V purely because the majority of their IT resources are remote and they want to save the cost of having to visit each server to capture its configuration and port it to a virtualization host. Others value the level of automation which allows them to accelerate how quickly they can migrate underutilized infrastructure to a more efficient virtualization host.

The speed of the migration determines how many you can do over a defined period of time. For a standard 100mbs network, a typical Windows server can be fully converted into a VM in under an hour assuming sufficient network access. The version of the OSdoes not really factor into the overall speed although Windows NT reboots one or two more times than Windows 2000 or 2003 as part of how the final configuration is applied.

The overall time savings when comparing PowerP2V to most any alternative is very large. The typical alternative for a Windows server would be to remove hardware dependencies using SYSPREP, capture the image to an off-line store, create the empty VM with the desired configuration, restore the image to the VM, undo the SYSPREP changes, boot the VM and manually finish the configuration changes. A typical manual conversion takesthe better part of the day for small servers, and perhaps two or three days for larger servers or more complex server configurations. With PowerP2V you can perform several concurrent conversions in a single day while reducing the risk of failure by several factors given that no manual intervention is required.

VI: One of the most problematic things of P2V migration is legacy boxes downtime during conversion. PowerP2Vbe able to convert a physical machine without interrupting the work? If so how will be done?

SP: In its current form, the source server is taken off-line to capture a complete instance for the purposes of migration. We are working to reduce the downtime which will allow a wider set of application for the technology. For example,for standard consolidation, most customers seem to be able to plan a maintenance window to implement the conversion. When PowerP2V is used for high-availability applications, the downtime needs to be minimal. Were committed to evolving PowerP2V in that way and have worked on some alternatives in our lab.

PlateSpin is shortly releasing a new flexible image format as an interim step on the conversion process. This will allow P2I and I2V conversions to be done which will help reduce the overall downtime by separating the steps of capturing the source from the step of creating the destination. This new feature, part of PowerP2V 4.0, will also add to the options customers have for provisioning new virtual machines, with the added advantage of the configuration flexibility currently enjoyed when doing a P2V directly.

VI: Another big problem of P2V migrations are hardware RAID configurations on machines to be converted. PowerP2V is able to manage such a configuration? If not what workaround can be used to migrate?

SP: PowerP2V captures a logical view of the server and therefore is not affected by RAID configuration details. Unlike the solutions that use physical imaging, we have not had any customer scenarios where the specifics of how the disk is configured affected the migration success. We can issue an update to PowerP2V normally within 2 to 5 business days to deal with hardware dependencies should they occur. In some cases, the customer can update PowerP2V for new hardware drivers using some field installation utilities we ship with the product.

VI: A really desired feature of this virtualization market segment would be preparing a perfect virtual machine on a lab environment and deploy it on a wide corporate desktop population, dramatically reducing installation task time. And nobody offered virtual to physical (V2P) migration till today. PowerP2V is going to become a PowerV2P also?

SP: One of our key messages announced during the recent VMworld User Conference in Oct, 2004 in San Diego, is the introduction of a PowerV2P capability. We previewed the initial implementation in our booth and will be releasing this shortly for trial.

We find V2P specifically important for customers using virtual infrastructure in a test lab and a mixed environment (or physical only) in production. The use of P2V and V2P allows for round-tripmigrations which is very valuable for the typical way in which servers are managed through their lifecycle.

VI: Actual virtualization players, EMC/VMware and Microsoft, offer products a company could buy together for different purposes and deployments. So IT managers start to feel the need of a conversion tool doing so called virtual to virtualconversion (V2V) from a virtualization platform to another. PowerP2V is able to convert to/from all VMware (ESX Server, GSX Server and Workstation) and all Microsoft (Virtual Server and Virtual PC) products?

SP: PowerP2V provides a full V2V capability to allow customers to implement multiple virtualization technologies and move server environments between them. This is in our currently shipping product.

VI: Microsoft has just released its own P2V migration tool, Virtual Server Migration Toolkit (VSMT), which is completely free. Where PowerP2V is different from it and why these differences would worth the price?

SP: As with all product comparisons, they are rooted in customer-specific requirements. One of the main differences is that PowerP2V does not require any specific 3^rd party solutions. The VSMT is integrated with ADS provisioning which means customers have to support and deploy ADS as part of their virtualization infrastructure. Many customers already have a multi-platform provisioning solution in place, so ADS is not a viable alternative. Secondly, PowerP2V provides native support for multiple VM architectures, something the market seems to be embracing. Finally, we offer substantive levels of automation and configuration flexibility, which we feel is the real value of PowerP2V in terms of being able to use it to properly optimize a data center over time.

VI: PowerP2V is able to virtualize even Linux boxes. Butsome IT managers already feel the need to convert *nix boxes (mainly SUN Solaris for x86 and BSD operating system family). This need could raise much when upcoming Solaris 10 will start to spread. And VMware is starting to support it as guestoperating system. PlateSpin is going to support any *nix box conversion?

SP: Yes, PlateSpin continues to work on extending Linux support to other UNIX operating systems. We will keep pace with customer requirements as they emerge. We currently cover most Red Hat and SUSE releases and are adding a few more Linux platforms over the next three months.We have seen a growing need for Solaris I86 support, not as much for FreeBSD. The approach we have taken to the development of our product makes it easy to extend as needed to support new platforms. If customers have a large need for a new platformto be certified, they should contact us directly.

VI: A third virtualization player is just appeared, Serenity Systems International, offering a new virtualization platform: Virtual Station (SVISTA). PlateSpin plans to support SVISTA in future PowerP2V releases?

SP: This is under consideration.

VI: PowerP2V greatly simplified legacy machines conversion, even if there are some issues, like moving from physical IDE storage system to a virtual SCSI one, an IT manager should know before starting migrations. PlateSpin plans to realize training courses about PowerP2V? If so Partner will be able to teach them?

SP: We have built a rapidly growing network of value-added partners across the globe. Many of these partners are now able to provide value-added services when working with PowerP2V, including helping the customer implement a complete server consolidation program. Customers should contact us for more information about our partner program as well as who to talk to if they are looking for value-added services for our products. We are also shortly launching a customer training program look for details on our web site in the coming weeks.

VI: Can you give us some more anticipationabout future technology features PlateSpin is developing for PowerP2V and when well see them?

SP: Our vision is to bring full OS portability to the data center. As such, we are workingto enhance PowerP2V to provide multi-directional conversions to all the relevant data center architectures based on IA technology. It will support not only adoption of virtual infrastructure but improve the way in which basic server management occurson traditional physical computing models as well.

We are also working on a series of related products for introduction throughout the year that enhance the IT specialists ability to optimize the data center, consolidate the data center and prepare a proper disaster/recovery program. All of these new products leverage the technology we have already built.

Our next release, scheduled for January 2005, introduces the support for flexible images mentioned above. It also adds additional Linux platforms, support for foreign language OS, and email notification, to make it easier to track a conversion through the job.

We also emphasize the needs of the systems integrator in the way we package and license our products. As such, we will be adding more features that support their unique needs along with the day-to-day operational requirements people have to get more out of IT technology.