SaaS data loss: The problem you didn’t know you had

Software giant Oracle has announced its purchase of server virtualization wannabe Virtual Iron, confirming months of rumor.

Having already grabbed the open source Xen hypervisor as the basis of its free-standing Oracle VM hypervisor, Oracle did not need Virtual Iron for its own VT and AMD-V dependent implementation of Xen. But Virtual Iron has some expertise and management tools that Oracle needs if it hopes to compete against Red Hat, Microsoft, Citrix Systems, and VMware as they are delivering better server virtualization tools.

Katana's VFe actually created a single image out of clustered machines, and it is debatable how well the software worked. It just so happened that VFe had its own virtualization hypervisor, which allowed a partition on this cluster of machines to be as small as 1/16th of a core all the way across all 16 cores in the system, with a maximum of 128 partitions per clustered systems.

This all sounded very cool - but InfiniBand didn't take off, the dot-com bust hit, and the VFe hypervisor wasn't compatible with VMware's ESX Server and the emerging alternative, the Xen hypervisor. And so Katana changed its name to Virtual Iron, dumped InfiniBand system clustering and its own VFe hypervisor, and transformed itself into a provider of a Xen hypervisor with an alternative set of system tools derived from the VFe stack.

By September 2005, Virtual Iron had raised $28.5m in three rounds of venture funding, backed by Intel Capital, Goldman Sachs, Highland Capital Partners, and Matrix Partners. Intel also agreed to help Virtual Iron tune its software for the on-chip VT virtualization features in its Xeon processors.

With the money and a revamped Xen stack, Virtual Iron tried to set itself up as a cheaper alternative to VMware and a better Xen implementation than the one that XenSource, the company behind the Xen project, delivered to the market. But once VMware went public, Citrix Systems bought XenSource, and Microsoft got its freebie Hyper-V into the field, Virtual Iron - even with its $65m in total venture funding - was just not up to the task of taking on these three companies. It has been very quiet in the past year after being very aggressive for the prior three.

Oracle put out a statement about the Virtual Iron deal, saying that until the transaction closes Virtual Iron will operate separately; it did not say how much it paid to acquire Virtual Iron, either. In a presentation on its website, Oracle said that it is reviewing Virtual Iron's product roadmap, and will eventually give Virtual Iron's customers guidance on product roadmaps and any changes in features or timing in the wake of the deal.

Oracle also wanted to be very clear that the product roadmap information provided by Virtual Iron in the past or Oracle in the future "does not represent a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decision." It expects the deal to close this summer, and at that time it will talk about how the Virtual Iron tools will be woven into its Oracle Enterprise Manager tools. In the meantime, Virtual Iron's tools are still being sold and supported.

The main bits of the Virtual Iron tool stack that Oracle was interested in are CPU and storage capacity management and power management, all features that Citrix and VMware have focused on in their latest hypervisor stacks (Essentials for XenServer 5.5 and vSphere 4.0, respectively, and neither shipping yet.)

Virtual Iron had also created a set of scriptable APIs that allowed system administrators to link the management of their Xen VMs into their other system management frameworks so physical and virtual machines can be orchestrated from a higher level and with the same tools. ®

Steps to Take Before Choosing a Business Continuity Partner