Nutanix, a cloud hardware start-up that's offering a hybrid scale-out compute-cum-storage appliance backed by $72 million in VC funding only half of which is reportedly spent, has put out next-generation software-defined data center products.

It's updating its server hardware and its software to deal with divergent workloads. It's going to a quad-node box made by Quanta and should be able to support 400 VMs per chassis, up from 300.

It's got VM-centric disaster recovery, adaptive compression and a new highly configurable hardware platform. The widgetry includes Nutanix OS 3.0 and NX-3000 series hardware. It's supposed to help enterprises build next-generation software-defined data centers.

Besides VM-level disaster recovery and adaptive post-process compression, Nutanix OS 3.0 delivers dynamic cluster expansion, rolling software upgrades and support for KVM, its second hypervisor after VMware's.

Its software enhancements, coupled with the configurable NX-3000 series platform, enable flexibility, performance and scalability in enterprise data centers.

With NX-3000, Nutanix delivers a configurable platform in which compute- and storage-heavy nodes co-exist in a single heterogeneous cluster. It includes hardware models that vary in capacity and the number of PCIe-SSDs, SATA SSDs and SATA DDs server nodes.

The nodes can have different CPU cores per socket and variable memory capacities. This allows for independent scaling of compute and storage in a single system that's optimized for every use case and can scale to address evolving business requirements.

The Scale-Out Converged Storage (SOCS) virtual disk controllers that make the Nutanix server cluster into a SAN so compute and storage are on the same cluster and the compute jobs are close to the storage. Nutanix uses Flash

The NX-3000 uses Intel's Sandy Bridge chips - the eight-core E5-2660 processors running at 2.2GHz - and delivers VM density in a 2U form factor.

Nutanix claims to be the first to deliver RAID, high availability, snapshots and clones at the VM-level.

It says it's implemented a highly differentiated VM-centric disaster recovery engine.

The new Nutanix OS 3.0 includes native storage-optimized disaster recovery that enables multi-way, master-master replication supposedly never seen before in traditional storage arrays.

Administrators can configure disaster recovery policies that specify protection domains and consistency groups in primary sites, which can then be replicated to any combination of secondary sites to ensure maximum business resiliency and application performance. And any Nutanix cluster can serve as both a primary and secondary site simultaneously for different protection domains, providing even more flexibility and choice.

Nutanix OS 3.0 is supposed to deliver best-in-class runbook (failover and failback) automation that's hypervisor-agnostic, which means native disaster recovery capabilities are available and consistent regardless of the underlying virtualization platform or management tools.

One of the pillars of the Nutanix solution is a highly efficient MapReduce-based framework that implements information lifecycle management in the cluster to achieve tiering, disk rebuilding and cluster rebalancing.

It's supposedly the first of its kind in the storage industry.

The same framework is being leveraged to deliver adaptive post-process compression of cold data as it migrates to the lower data tiers, so as not to impact the normal IO path.

By leveraging the information lifecycle management capabilities inherent in Nutanix' software, the system dynamically determines which data blocks to compress based on how frequently they're being accessed by the VMs.

Post-process compression is ideal for random or batch workloads and delivers the highest possible overall performance. In addition, Nutanix' OS 3.0 supports basic in-line compression that works as the data is being written, which is better suited for archival and sequential workloads.

The company says, "While our existing storage solutions support compression in general, the granularity of Nutanix compression allows us to set policies at the VM level, ensuring maximum business value and storage utilization,"

With Nutanix OS 3.0, the company is supposed to deliver on its commitment to bring all of its enterprise features to the broadest range of platforms in the industry.

The software, which was designed to be hypervisor-agnostic, will now support KVM and VMware vSphere 5.1.

Regardless of the underlying virtualization platform or management framework, enterprises benefit from all of the capabilities of the Nutanix software.

The KVM hypervisor provides financial flexibility for enterprises and works well in workloads such as Hadoop.

Nutanix OS 3.0 also uses a discovery-based protocol to auto-detect new nodes added to the same network as a cluster, enabling administrators to quickly and easily expand a cluster without incurring any downtime.

In the background, the system will then rebalance the data across the entire storage pool, including the newly added nodes, to provide maximum I/O performance.

The new software also uses software-defined networking tricks to achieve rolling software upgrades in the always-on cluster. Upgrades are delivered in a peer-to-peer framework to enable rapid software upgrades while retaining maximum cluster availability.

The features and capabilities delivered in Nutanix OS 3.0 and NX-3000 are supposed to usher in a new era of business resiliency and data center optimization.

The start-up thinks it's displaced $25 million in server and SAN storage sales and is close to doubling sales every quarter. Its co-founder and CEO Dheeraj Pandey built the first Exadata clusters at Oracle. Co-founder Mohit Aron was chief architect at Aster Data and lead designer of the Google File System that led to Hadoop.