Nutanix has the potential to replace most of today's traditional storage solutions. These are classic hybrid SAN arrays (dual and multi-controller), NAS Filers, newer All-Flash Arrays as well as any object, big data etc. use cases. For capacity, it usually comes down to the price for large amounts of data where Nutanix may offer higher-than-needed storage performance at a price point that isn't very attractive. This has been addressed in the first step using storage-only nodes which are essentially intelligent disk shelves (mainly SATA) with their own virtual SDS appliance preinstalled. Storage nodes are managed directly by the Nutanix cluster (the hypervisor isn't visible and no hypervisor license is necessary). While this is going in the right direction, larger storage nodes are needed to better support "cheap, big storage" use cases. For typical big data use cases today's combined compute and storage nodes (plus optionally storage-only nodes) are already a very good fit! The Nutanix File Services (Filer with active directory integration) are a very welcomed addition customers get with a simple software upgrade. Currently, this is available as a tech preview to all Acropolis Hypervisor (AHV) customers and will soon be released to ESXi as well. This is one example of a service running on top of the Nutanix distributed storage fabric, well integrated with the existing management layer (Prism) offering native scale-out capabilities and One-Click upgrade like everything else. The demand from customers for a built-in filer is big, they are looking to not depend on legacy filer technology any longer. We are looking forward to seeing this technology mature and offer more features over the coming months and years. Another customer need is to be able to consume Nutanix storage from outside the cluster for other, non-Nutanix workloads. These could include bare metal systems as well as non-supported hypervisors (e.g. Xen Server etc.). This functionality (called Volume Groups) is already implemented and available for use by local VMs (e.g. Windows Failover Cluster Quorum) and will soon be qualified for external access (already working from a technical point of view including MPIO multi-pathing with failover). It will be interesting to see if Nutanix will allow active-active access to such iSCSI LUNs (as opposed to the current active-passive implementation) with the upcoming release(s). Imagine if you upgraded your Nutanix cluster (again this would be a simple One-Click software upgrade) and all of sudden you have a multi-controller, active-active (high-end) storage array. (Please note that I am not a Nutanix employee and that these statements describing possible future functionality are to be understood as speculation from my side which might never become officially available.)
