To hear advocates talk about NVMe – a de facto standard created by a group of vendors led by Intel to connect flash memory storage directly to a PCIe bus (that is, without using a SAS/SATA disk controller) – it is the most revolutionary thing that has ever happened in business computing.  While the technology provides a more efficient means to access flash memory, without passing I/O through the buffers, queues and locks associated with a SAS/SATA controller, it can be seen as the latest of a long line of bus extension technologies – and perhaps one that is currently in search of a problem to solve.

As I review the marketing pitches of many software-defined storage products today, I am concerned by the lack of attention in any of the software stack descriptions to any capabilities whatsoever for managing the underlying hardware infrastructure. This strikes me as a huge oversight. The truth is that delivering storage services via software — orchestrating and administering the delivery of capacity, data encryption, data protection and other services to the data that are hosted on a software-defined storage volume – is only half of the challenge of storage administration. The other part is maintaining the health and integrity of the gear and the interconnect cabling that provide the all-important physical underlayment of an increasingly virtualized world.

Personally, I am getting rather tired of the dismissive tone adopted by virtualization and cloud vendors when you raise the issue of disaster recovery.  We previously discussed the limited scope of virtual systems clustering and failover:  active-passive and active-active server clusters with data mirroring is generally inadequate for recovery from interruption events that have a footprint larger than a given equipment rack or subnetwork.  Extending mirroring and cluster failover over distances greater than 80 kilometers is a dicey strategy, especially given the impact of latency and jitter on data transport over WAN links, which can create data deltas that can prevent successful application or database recovery altogether.

At the end of March, an event little known outside of a small community of vendors, will happen:  World Backup Day.  Expect a flurry of blogs and tweets and posts and all of the other stuff that goes along with such marketing events.  Then, expect the discussion to go silent for another year…unless a newsworthy data disaster occurs.

Hyper-converged infrastructure, when we started to hear about it last year, was simply an “appliantization” of the architecture and technology of software-defined storage (SDS) technology running in concert with server virtualization technology. Appliantization means that the gear peddler was doing the heavy lift of pre-integrating server and storage hardware with hypervisor and SDS hardware so that the resulting kit would be pretty much plug-and-play.

It used to be that, when you bought a server with a NIC card and some internal or direct attached storage, it was simply called a server. If it had some tiered storage – different media with different performance characteristics and different capacities – and some intelligence for moving data across “tiers,” we called it an “enterprise server”. If the server and storage kit were clustered, we called it a high availability enterprise server. Over the past year, though, we have gone through a collective terminology refresh.