Microsoft built virtualization on top of its operating system – which supposedly gave it something of an edge – at least on the cost side. So now VMware, under serious competitive pressure because of its envied market leadership – and needing to show technological distance between its widgetry – at least its vision – and the copycats – has retaliated with an operating system of its own, a so-called Virtual Datacenter Operating System (VDC-OS).

Unlike a traditional OS, which is optimized for a single server and supports only those applications written to its interfaces, VMware says VDC-OS serves as the OS for the whole datacenter and supports any application written to any OS, from legacy Windows applications to modern distributed applications that run in mixed operating system environments.

This newfangled VDC-OS, when it’s fleshed out sometime next year, is supposed to pool all kinds of hardware resources – servers, storage and network – into an aggregated on-premise cloud – and, if needed, federate workloads to external clouds for even more compute capacity – sorta like Google and Amazon.

VMware says data centers running on the VDC-OS become highly elastic, self-managing and self-healing.

This is the way it will work.

First, it will deliver a set of infrastructure services called Infrastructure vServices to seamlessly aggregate servers, storage and network as a pool of on-premise cloud resources and allocate them to applications that need them most.

Second, it will deliver a set of application services called Application vServices to guarantee the right levels of availability, security and scalability to all applications independent of the operating system, development frameworks or architecture on which they were built to run.

Third, it will deliver a set of cloud services called Cloud vServices that federate compute capacity between the on-premise and off-premise clouds.

According to this new roadmap, VMware is still leaving the management of physical servers to third-party widgetry like OpenView and Tivoli.

The Application vServices widgetry will include Fault Tolerance, which VMware says will guarantee zero downtime and zero data loss to all applications by warding off hardware failures without the cost and complexity of hardware or software clustering solutions.

It will also offer Data Recovery for all applications via disk-based backup and recovery.

VMware figure that Fault Tolerance and Data Recovery coupled with its existing abilities to automatically restart virtual machines affected by software failures, minimize planned downtime and automate the failover of an entire site will make VDC-OS “the safest platform for applications.”

The company will also deliver services such as VMsafe, which it says will provide “X-ray vision” into virtual machines to detect and stop previously undetectable viruses, rootkits and malware before they can infect a system. Checkpoint, IBM, McAfee, Radware, Symantec and TrendMicro are supposed to deliver VMSafe-integrated products next year that protect VMware virtual machines better than physical machines or other virtualization solutions.

VMware’s also got a way to Hot Add virtual CPU, memory and network devices to virtual machines enable applications to scale seamlessly without disruption or downtime and support very large virtual machines with up to eight virtual CPUs, 256GB of RAM. Another service called VMDirectPath is meant to provide enhanced network and storage I/O performance for transaction applications.

The company will also have vApp and vStudio, capabilities that should make it easier for administrators to deploy and manage applications.

It says vApp will turn new or existing applications into self-describing and self-managing entities.

vApp leverages the Open Virtual Machine Format (OVF), a proposed standard, to specify and encapsulate all components of a multi-tier application as well the operational policies and service levels associated with it. vApp gives application owners a standard way to describe operational policies for an application that the VDC-OS can automatically interpret and execute.

VMware Studio is an authoring and configuration tool that will enable ISVs and enterprises to construct Virtual Appliances and vApps. Instead of deploying and managing operating systems and applications separately, VMware says a vApp can be deployed and managed as a single entity.

The new Infrastructure vServices in VDC-OS will extend the concepts of resource pooling to storage and network to enable elastic, on-demand allocation of on-premise resources. The new Infrastructure vServices are said to complement VMware DRS, which provides on-demand allocation of processing capacity and memory to applications. VMware says the VDC-OS is geared to minimize the resources consumed per application, significantly lowering the capital and operating costs per application.

Such services include vStorage Thin Provisioning and vStorage Linked Clones, which will reportedly reduce storage requirements by up to 50% and a vNetwork Distributed Switch, which will abstract the configuration of virtual networking from the host level to an aggregate data center level, simplifying setup and change of virtual networking and allowing the delivery of cluster-wide networking services. The switch will enable Network VMotion, the preservation of all network and security policies when a virtual machine is being migrated.

VMware says VDC-OS lays the foundation for more efficient and automated IT management processes. The combination of existing vCenter products and the widgetry due in 2009 are supposed to equip users with management capabilities to automatically provision new VMs and vApps, ensure compliance with established configuration standards, right-size every element in the environment, predict and manage capacity, and allocate the costs back to the business.

New management capabilities include vCenter ConfigControl, which will extends policy-based change and configuration management with automated enforcement across every aspect of the VDC-OS; vCenter CapacityIQ, which will continuously analyze and plan capacity to ensure optimal sizing of virtual machines, resource pools and the entire datacenters; vCenter Chargeback, which will automate tracking of costs and chargeback to the business enabling IT to function as a utility with true visibility into operating costs; vCenter Orchestrator, which will enable the development of customized workflows that automate operational tasks through a simple drag-and-drop interface, without scripting; and vCenter AppSpeed, which will automatically ensures application performance levels. It will monitor end-user response time for applications, correlate these response times with different elements in the infrastructure, and trigger remedial actions to alleviate bottlenecks.