Metacloud plays hardball.

It has no patience with pilots or proofs-of-concept. It'll only do production Infrastructure-as-a-Service installations, which is kind of uppity for a start-up that only hit the radar last October - even if it did come out of the closet with an unidentified Fortune 100 reportedly under contract.

It didn't even have a name for its OpenStack distribution until now. It's calling the private cloud it's built on the latest Grizzly version of OpenStack Carbon|OS.

See, Metacloud's founders are pretty sure of themselves because they figure they've done this kind of thing before. When President Steve Curry was at Yahoo for over a decade he managed its global storage operations that handled hundreds of petabytes of content and user data. And when CEO Sean Lynch was at Ticketmaster, the third-largest e-commerce system in the world, he ran its technical operations.

The company claims its operational experience is "a unique asset, found nowhere else in the private cloud services sector today."

Anyway, Lynch and Curry - who give interviews together - say Carbon|OS combines the simplicity of the public cloud with the performance, security and cost advantages of a private cloud - largely because it runs on clients' existing hardware infrastructure but at supposedly as much as a whopping 85% utilization rate because of OpenStack's multi-tenancy design.

That's utilization way beyond what companies are reckoned to be getting either pre- or post-virtualization.

And, if not that, then because it's managed and supported remotely 24/7 by Metacloud's people; the customer retains control of its virtual servers, data and intellectual property, consumes compute, and passes any heavy lifting off to Metacloud, which claims to be doing this for some of the nominally best and brightest.

Lynch and Curry say they've invested Carbon|OS with a number of breakthrough features that make it ideal for demanding enterprise environments and superior to other public and private clouds services.

These features, some of which existed back in October on earlier iterations of OpenStack, include:

• True self-service: The Carbon|OS invite-based tenant management system lets end users access the compute resources they need when they need them without having to involve IT administrators. They simply leverage their existing Active Directory credentials to access compute power and collaborate with co-workers instantly. It's supposed to enhance enterprise agility and help eliminate the key drivers of shadow IT by removing the obstacles users traditionally encounter when procuring IT resources in-house.
• Advanced instance scheduling & placement: Carbon|OS offers dozens of new algorithms that let administrators put VMs wherever they want for maximum performance. Latency-sensitive applications can be housed on VMs that are right next to each other. Large Hadoop clusters can be broken up into smaller nodes that have decreased network and I/O latency. And web servers can be spread across different racks throughout the data center for maximum redundancy.
• Enhanced dashboard: Previously many OpenStack capabilities were accessible only via API and CLI tools, which was fine for experienced administrators. The improved dashboard in Carbon|OS provides a more intuitive, easier-to-use environment that will appeal to a larger population of IT professionals.
• Horizontally Scalable Control Plane: Typical disaster recovery systems require a complicated set of technologies to do the right thing under the worst conditions. Carbon|OS is arranged differently; it runs cloud orchestration services across all nodes in the control plane all the time so the system is continually tested. This not only decreases interruptions during unplanned outages, it improves performance by utilizing all the resources allocated to the control plane.
• Commodity-based block storage: Carbon|OS delivers block storage capabilities through Ceph, the open source solution that lets IT departments pool storage across commodity compute nodes. This mitigates the need for expensive shared storage hardware and frees organizations to purchase commodity equipment from any manufacturer offering the best price.

Metacloud's boys say enterprises are struggling with hard-to-manage scale-out clouds and can spend weeks, if not months, provisioning Amazon widgetry because they "don't have the bench to operate" the stuff. There have also been unconfirmed reports of OpenStack losing design-wins.

Metacloud claims to have made the cloud easy to manage.

It reportedly has more than 10 customers now including one Fortune 100 and all its subsidiaries that it snared away from VMware.

Metacloud promises to cut TCO in half and says it backs up its services with SLAs "with teeth."

Carbon|OS will work on any x86 servers and sells for a fixed monthly subscription amount based on the number of CPUs used. Updates and new features are delivered as a SaaS.

The company, which is targeting a market worth tens of billions of dollars and expects to be operating as a global company with a $100 million revenue run rate by 2017, is backed by some secret amount of money from Storm Ventures and Yahoo co-founder Jerry Yang, proving it pays to have rich friends.