Um, where's the Trump angle?
The article looks like you wanted to ride the Trump hate wave, but realized you have no idea how to throw innuendo in there and gave up, but didn't change the title, hoping for a few clicks.
Block-based storage (and particularly that based on Fibre Channel) has always been the favoured child when it comes to vendors talking about their products and solutions. All of the high-end performing devices tend (but not exclusively) to be block-based. However we’re seeing an increased focus on high performance NAS …
The irrelevant title is probably more pathetic click-bait, so I'm adding a filter to trash titles like.
If this is not just click-bait, the author/editors need to get a clue that Trump seems a much lesser evil than the obviously very ill, corrupt and criminal H. Clinton.
...well more correctly IBM Spectrum Scale Storage, is a block based protocol (unless you're using the built in NFS bridge), putting the onus of working out where the storage for files is onto the client.
If you're taking about it working like a NAS, then you've probably come across it in it's SONAS storage appliance persona, not in it's GPFS client/server software defined storage persona.
The high latency of SAN compared to SSDs means block storage users like databases are slowly but inexorably moving to shared-nothing, direct attach (DAS) architectures like PCIe and NVMe, with the networking being handled at a higher level by the database itself. Aerospike and Spark are good examples of this trend.
NAS is the only way storage vendors can still peddle expensive and increasingly irrelevant storage arrays. The other is acting as a premium backend for virtualization (i.e. vMotion), and that's being eaten up by software as well.
Fazal Majid talks of high latency SAN but mine is consistently sub-1ms. This is more than rocket fast for all the applications most mortals run. Real time trading platforms might need more but not most things. The speed potential in things like the NVMe fabrics is interesting, but irrelevant for 99% of use cases. SANs will be with us for a while yet, switching to all flash for most production data over the next 18 months or so, but they will still be with us in 2, 3, 4 and more years.