virtualization.info

So far we have dedicated a lot of space to Oracle, in terms of what virtualization offering it could provide and what mistakes may compromise its presence as a relevant player.

The Sun acquisition has not closed yet, so the company cannot disclose any specific plan. Without concrete information about that, what we have published so far, and what follows below, is pure speculation.
Nonetheless it’s worth spending some more time evaluating the strategy that Oracle may put in place and how it may impact the current players.

As already said many times, now the company is in the unique position to offer an entire computing stack, including servers, storage, the hypervisor, the operating system, the middleware, some of the most used business applications, thin clients, a VDI connection broker and an enterprise management software to coordinate all of the above.
Leveraged in the right way, and assuming Oracle may become a credible virtualization player, it represents a remarkable competitive advantage for some customers (while others can clearly see it as a painful way to lock themselves in).

VMware, Citrix, Microsoft and now Red Hat, have to deal with multiple vendors and support thousands of different hardware and software components (VMware just launched a certification program for software. Why did they have to do that?). And a lot can go wrong when your hypervisor is the glue that keeps together servers, storage, network, guest operating systems, enterprise management agents, guest middleware and guest applications.
Oracle is the only one, in the virtualization market, that could say, “We know exactly what happens at every level of the stack, because we provide all the components; we can guarantee the behavior and the performance of our virtual infrastructure because there are no 3rd parties involved.”

There’s another company that is in a similar position, but in a completely different market: Apple.
Apple develops its software and its systems, and is fully in control. Steve Jobs considers this one of Apple’s biggest assets:

We're the only company that owns the whole widget -- the hardware, the software, and the operating system. We can take full responsibility for the user experience. We can do things that the other guy can't do.

It is a lock-in, the growing number of issues around the iPhone App Store approval process confirms this, but it’s a huge success.

Of course the consumer market and the enterprise market are different worlds, but Oracle may well pitch its virtualization offering in the same identical way.
If so, Oracle is going to compete with the just born Virtual Computing Environment (VCE) coalition, a nice acronym that also means VMware Cisco EMC, the three companies that founded it.

The value of VCE products, the self-contained virtual data centers called Vblocks, is not only in the hardware and software that make the units. It’s in the fact that VMware, Cisco and EMC design, produce, test and certify the units to serve a specific amount of virtual machines, for a specific amount of users, interacting with specific workloads, that perform in predictable ways.

In other words the VCE coalition saves the customer the huge investment of designing his own data center and the costs of designing it in the wrong way.
When the customer buys a Vblock, he is not just buying the hardware and the software. He is also buying the know-how that these three companies put in the machines. A know-how that he would have to produce by himself or buy somewhere else.

To validate this approach, VMware Cisco and EMC had to form a new entity and share investments, because none of them controls the full stack. Oracle does, and if the future of IT will be dominated by modular data centers, where a single vendor provides self-contained units that customers just stack up together, then Oracle now has the opportunity to become a leader in that future just as much as Cisco.

The difference between Cisco and Oracle is that the former has already clarified its interest in doing so and took several steps to change its current image of networking provider, while the latter… well, the latter still is the well known database giant. And no more than that.

Labels: Cisco, EMC, Oracle, VMware

In the last two years Cisco made at least two long-term key investments in the server market: invested over $150 million in VMware and became a player with its own blade system Unified Computing System (UCS).

Cisco wants to sell and interconnect next generation data centers. To do so it needs servers, storage, networking, software abstraction and software management.
EMC is helping with storage and software management, VMware is helping with software abstraction.

The three worked together for some months and then announced a formal coalition, that will sell these integrated data centers through channels and direct relationship with customers (through a company called Acadia).

The joint effort of these three companies may be just an elaborate (yet remarkable) marketing exercise, but it certainly had an impact on the market.

It may have forced IBM to attempt the acquisition of Sun (which ultimately went to Oracle).
It may have forced HP to buy 3Com for $2.7 billion.

HP is the company that has the most to lose because of the VMware-Cisco-EMC coalition considering that it currently sells 36% of all physical servers that will run virtualized workloads, and that it’s the leader in this segment, ahead of Dell and IBM.

Cisco may use VMware to slip into the enterprises where HP reigns, like it attempted to do on the VMworld 2009 data center.
If Cisco becomes a real threat, HP will need servers, storage, networking, software management and software abstraction to counter it.
The question is not if HP wants or not to adopt a strategy that is similar to the Cisco one. The question is how long before the Cisco plans will allow HP to play nice with all three major virtualization vendors instead of buying the only one it can?

Labels: 3Com, Acquisitions, HP

At the end of August, Citrix announced a new major effort around Xen and cloud computing to counter the release of VMware vCloud Express.

The details of this project were scarce at that time and beyond the name, Xen Cloud Platform (XCP), and the intent to integrate new and existing technologies, Citrix didn’t disclose much more.

Now the things are getting cleaver, with the Xen.org entity detailing the list of proposed components for XCP 1.0 and makes available the platform for download:

• Latest Xen 3.4.1
• Linux 2.6.27 Kernel
• Windows PV Drivers, Microsoft Certified (Binary Only)
• XAPI Enterprise-class Management Tool Stack (web based management interface)
• VM Lifecycle: Live snapshots, checkpoint, migration
• Resource Pools: Safe live relocation, auto configuration, DR
• Host Configuration: Flexible storage management, networking, power management
• Event Tracking: Progress, notification
• Secure Communication using SSL
• Upgrade and Patching Capabilities
• Real-time Performance Monitoring and Alerting
• Basic SR-IOV Support
• CDROM and Network Host Installer
• Full Featured “xe” CLI and web services API

Xen.org also published a tentative roadmap for version 1.0:

• vSwitch Integration - first step to enabling multi-tenant network infrastructure, to enable firewall and routing rules to follow VMs as they migrate, and to enable flexible traffic monitoring of virtual ports
• Netchannel 2 Integration - improve scalability of xen networking on larger systems and to accelerate inter-VM traffic
• SR-IOV Networking - Although Xen support SR-IVO NICs today, configuration requires manual steps. By extending the control strack we can make SR-IOV simply a transparent optimization that is enabled automatically where possible
• Booting guests from SR-IOV HBAs
• Libvirt bindings
• Native support for OVF in the tool stack
• Drive DMTF standards for virtualization and cloud
• Smart error recovery to minimize impact of hardware errors
• Work closely with other projects and vendors to enable web-based mutli-tenant mgmt and provisioning; e.g. Eucalyptus, Enomaly, OpenNebula, etc.
• Increased management scalability for dealing with 1,000s of Xen hosts - federation of resource pools
• Aggregation of cheap local storage - integrated drdb/parallax
• OCFS2 integration

Labels: Citrix, Xen

Last week, at its TechEd Conference in Berlin, Microsoft published updated roadmaps for most of its products.
The most interesting one from a virtualization perspective is the one about System Center family.

It seems that Virtual Machine Manager (SCVMM) will not receive any major update before 2011:

Thanks to Bink.nu for the news.

Labels: Microsoft

Five months after the first beta, Microsoft is ready to push out the beta 2 of Visual Studio Team System 2010 Lab Management, a special version of the popular IDE that interacts with Hyper-V R2 and System Center Virtual Machine Manager (SCVMM) 2008 R2 to provide a fully featured virtual lab automation platform.

There’s not really much to say about this new beta, expect reporting a few improvements in the setup and administrative GUI, along with support for network fencing with virtual machines that are acting as domain controllers (this last one is a very welcome addition).

.

The Visual Studio Lab Management team is publishing a number of in-depth walk-through about how to use the platform and how special features (like network fencing) workIt’s really worth a check.

Labels: Microsoft, Virtual Lab Automation

Last week VMware finally released the much awaited View 4.0, which supports vSphere 4.0 and introduces the software-only version of the Teradici remote desktop protocol PCoIP.

VMware is offering two versions of View 4: Enterprise (which includes vSphere and View Manager 4.0), priced at $150 per concurrent user, and Premier (which also includes View Composer and ThinApp), priced at $250 per concurrent user.

Of course the key aspect of this release is how well PCoIP performs on LAN and WAN scenarios.
Unfortunately the product will be available for download on November 19, so for now it’s impossible to make a performance analysis and comparison with Microsoft RDP 7, Citrix ICA/HDX and the other tens of  alternatives that are flooding the VDI market.

The major problem with PCoIP is if its performance is so great to justify the adoption of a new proprietary remote desktop protocol at its 1.0 release (the protocol is more mature than that but so far relied on hardware components).
Many customers may want to be careful here, mostly considering that VMware and Teradici just have a co-development agreement, which is not even exclusive.
What happens if Teradici is acquired by a VMware competitor or if the company suffers major issues?
And most of all, what happens if one year from now VMware consider this protocol unpractical and too expensive to optimize and decides to replace it, for instance, with the just ratified Net2Display standard? 

Anyway a lot has been already said.

Brian Madden already published a brief FAQ list, which includes a couple of interesting details:

• The PCoIP client only supports Windows at the moment. Linux and Mac OS versions are expected next year
• View 4.0 will fully support Microsoft Windows 7 as guest OS in early 2010

Chad Sakac already published a blueprint to design a View 4.0 architecture with the recently announced VMware/Cisco/EMC hardware called VBlock.
The solution (a VBlock 1) fits over 2,048 virtual desktops and costs $750 per seat all inclusive:

The paper includes some performance analysis. It doesn’t clarify if the numbers are obtained when using the RDP or the PCoIP protocol (assuming this will make any difference) but it’s really worth a check.

Labels: Releases, VDI, VMware

Last week Lanamark, another startup that, like VKernel, could be impacted by the launch of VMware CapacityIQ, released Suite 2009 R2.

And like VKernel, Lanamark is looking around, introducing support for Windows Server 2008 R2 Hyper-V, along with refreshed support for VMware (now up to vSphere 4.0) and Citrix (now up to XenServer 5.5).

Suite 2009 R2 also introduces enhancements to its online dashboard, but the most important thing is that the product can now collect data from up to 50,000 systems.

At the moment the company only offers a hosted version of its platform, but if VMware CapacityIQ will start to get some traction, Lanamark may be obliged to reconsider its go-to-market strategy and give the product on-premises.

Labels: Capacity Planning, Lanamark, Releases

Just a few weeks after the release of VMware Fusion 3.0, Parallels answers with Desktop for Mac 5.0.

The company released the previous edition of this hosted virtualization platform exactly one year ago.

The new version 5.0 focuses on performance, claiming a 300% improvement on virtual machines operations (like start/stop a virtual machine) and up to 22% faster performance compared to Fusion (the study was conducted by Crimson Consulting Group but there’s no documentation for that).

Beyond that, the product ships features that seem on par with the ones provided by Fusion 3.0, with some advantages here and there:

• Support for 64bit
• Support for 8-way vCPUs
• Support for Aero (Windows 7/Vista), OpenGL 2.1 (Windows XP and Linux) and DirectX 9.0c with Shader Model 3
• Support for Mac OS X 10.6 codename Snow Leopard (32bit and 64bit, host and guest OS)
• Support for multi-touch gesture in Windows guest applications
• Virtual Machine Lock Down

At this point there’s not much differentiation between the Parallels Desktop and VMware Fusion. It’s primarily a matter of personal taste and capability to promptly solve the issues that demanding Mac customers will point out.
It will be very hard for Parallels to keep its market share unless they releases some major exclusive features.

Labels: Parallels, Releases

At the end of October VMware released its first capacity planning tool: CapacityIQ.

As always happens when a market leader expands into a new segment, its previous partners suddenly turn into competitors.
Some of them are able to keep the pace and deliver value on top of the new product, but this requires a number of resources that not every startup has, and a cooperation between the two companies that the newcomer may be uninterested in.

VMware is in the capacity planning space since ever with its hosted and free service Capacity Planner. But so far the product has been accessible only to the Professional Services Organizations (PSOs), which granted enough room to the partners to sell their on-premises products.

Now, with CapacityIQ around, VMware partners may lose market share.
They still can try compete on price, number of features and frequency of updates, but will always be customers that want everything from a single vendor, and there will always be vendors that gives away their newest product to establish a presence in a new segment.

VKernel is one of those partners: the company criticized the value of CapacityIQ at launch, but now its running to give away its flagship product for free (for a limited timeframe anyway).
If you download their Capacity Analyzer before the end of the year, you’ll receive a free license for unlimited sockets. Perpetual.

And of course VKnernel now is also in hurry to clarify that support for Microsoft Hyper-V and Citrix XenServer is coming.

Update: It seems that VKernel is not really giving away its flagship product, but a limited version called Capacity Modeler. Possibly the same tool that the company launched as beta in October 2008.
This post will be further updated with details as soon as possible.

Labels: Capacity Planning, VKernel

Yesterday, finally, Red Hat announced the availability of its new virtualization offering, which includes a platform based on KVM and an enterprise virtualization manager.

The company already released Enterprise Linux (RHEL) 5.4 in mid September, which features KVM in the same way (despite technical differences in the architecture) Microsoft Windows Server 2008 features Hyper-V.
The problem is that RHEL 5.4 plus KVM may be not enough to compete against lightweight, dedicated platforms like VMware ESX and Citrix XenServer. Additionally, RHEL 5.4 lacks of enterprise management tools that customers can use to control large scale virtual data centers.

This gap is filled today with the release of Enterprise Virtualization Hypervisor (REVH) and Enterprise Virtualization Manager for Servers (REVMS).

REVH is a stripped down version of RHEL 5.4, with the following characteristics (partial list):

• Support for Intel VT / EPT and AMD-V / RVI
• Support for up to 64 physical CPUs (up to 256 core)
• Support for up to 1TB physical RAM
• Support for up to 16 vCPUs
• Support for up to 64GB vRAM
• Support for memory overcommit (page sharing only, depending on Linux Kernel Same-page Merging)
• Support for physical NICs bonding and multipath I/O
• Support for NFS, iSCSI and Fibre Channel
• Support for RHEL (from 3 to 5) and Windows (2003, 2008 and XP) guest operating systems. 
  For Windows guests Red Hat offers paravirtualized (network and block) drivers based on the VirtIO standard, which are certified by Microsoft thanks to the SVVP certification.

Red Hat reports that KVM can handle up to 600 virtual machines within a single host.
Its platform, based on KVM, is reportedly able to handle more than 400 virtual machines within a single host (with 32 cores and 1TB physical RAM).
The company also claims that it can reach up to 95% of real hardware performance for mission critical workloads like SAP or Oracle Database.

REVMS instead supports the following capabilities:

• Virtual machines live migration (across NFS, iSCSI and FC shared storage)
• Virtual machines high availability (if a host dies all its virtual machines are restarted on another one within the same cluster. It requires an out-of-band management interface such as IPMI, Dell
  DRAC, HP iLO, IBM RSA or BladeCenter for host power management.)
• Virtual machines dynamic resource management (storage, networks and computing capability can be aggregated in resource pools. The System Scheduler relocates the VMs across the hosts that are part of the pool following the system policies and using live migration)
• Hosts maintenance mode (when the host is put in maintenance REVMS uses live migration to move virtual machines elsewhere)
• Hosts power management (the System Scheduler can use live migration to relocate the VMs on low activity hosts and power down the unnecessary servers)
• Virtual machines thin provisioning (the REVMS component called Image Manager allows to overcommit storage
• Virtual machines snapshots (snapshots can be scheduled and used as recovery points)
• Virtual machines templates
• Role based access control and support for Microsoft Active Directory for the management console
• APIs

Funny enough, it seems that the REVMS console is only available for Windows clients (we’ll double-check with Red Hat on this and update this article accordingly).

The two products, bundled together with the name of Red Hat Enterprise Virtualization for Servers, are sold through a subscription model. Price starts at $499 for 1 socket with 12x5 support.
Of course REVMS can manage the KVM platform included inside RHEL 5.4, but the operating system must be purchased separately.

To justify the value of its new offering, Red Hat even prepared a feature comparison matrix which includes VMware vSphere 4 and VI 3.5, as well as Microsoft Windows Server 2008 R2 with Hyper-V, and a price comparison matrix that will generate endless discussions (we already know it):

The last piece of the new offering, Enterprise Virtualization Manager for Desktops (REVMD), which is the SolidICE product acquired from Qumranet in September 2008, will be released in early 2010.

That piece represents the real opportunity for Red Hat to attract new customers and compete with more mature competitors. Convincing customers to change their hypervisor of choice for server consolidation isn’t easy at all, but there’s still a huge, untapped opportunity around client consolidation (aka VDI) and enough confusion to give KVM plus SPICE a chance to gain some market share.

Another critical point is how the ecosystem will welcome this new offering. The Red Hat offering isn’t able to cover every need a customer may have around managing a virtual data center, so partners are critical.
The first one to jump on the Red Hat bandwagon is VMLogix, which announced its commitment to support REVH in its lab management solution Lab Manager.
Many others have to follow to make this offering a valuable alternative to VMware, Citrix and Microsoft virtualization platforms.

Labels: Red Hat, Releases