Why Nutanix is not like Stonefly or your typical VSAN appliance
Thanks for taking a look at Nutanix Complete Cluster, Chris, and hope we get a chance to show you the product in action soon. Maybe at vmworld?
We believe the comparison with Stonefly is not completely accurate, so wanted to share some more detail.
Stonefly seems to simply consume a server's worth of local PCIe SSD and then serve it up via an iSCSI target. In contrast, Nutanix provides a fully redundant , scale-out, clustered solution. That means that storage from PCIe SSD and accompanying SATA drives from all nodes across the cluster are aggregated and presented as a single name space & iSCSI target – what we call Scale-Out Converged Storage (SOCS). High performance comes from the VM data being constantly optimized to be local and simultaneously replicated and distributed across all nodes in the cluster for redundancy.
The Nutanix architecture leverages distributed systems techniques pioneered by Google (we have key architects, eg our CTO, who worked on GFS and Oracle Exadata), and can thus scale to hundreds of nodes seamlessly in a building-block architecture. Growing storage and compute resources is as simple as adding another node. With Stonefly, it seems that only one server is providing the data, and if that server dies, then down goes the entire population of VMs. Perhaps they can provide an active-active solution, two nodes, but they'd be limited in the long term in terms of scalability.
Feature-wise, we're providing enterprise-class functionality like HA, DR, thin provisioning, HOT (heat-optimized tiering), VM-based policies, backup and more to be announced soon, so to be clear, this is not meant to be an SMB solution, but rather a simpler, more scalable, and dramatically lower cost alternative to vBlocks/FlexPods, for enterprise virtualization use cases.
BTW, we are aiming to be 100% channel driven, so interested resellers should contact us here: http://www.nutanix.com/partners.html.
- The Nutanix Product Team