The solid state of storage in 2018: Latencies, they are dwindling. On-premises, the kit is glistening... 2018 wasn't too shabby for storage as the three-pronged attack of IoT edge, AI and multi-cloud defeated notions of the public cloud eating on-premises storage. Revenues also rose, capacities increased and latencies dived to boot. Capacity rises followed new technology promising extra layers and QLC (4 bits/cell) for flash, and …
The glut of disk drive companies in the late 80's, early 90's when they were all funded with the proposition of getting a "large share" of the market. Now we have only a few true drive companies.
Will solid state/cloud/etc.. companies follow? I don't know, but I suspect that there might be a falling out in a bunch (technical term) of years.
Amazing what you can get VC money for these days.
Me? I will enjoy the prices going down as the competition heats up. That's a "good thing".
1) When it happens, and is fully supported, Storage Class Memory will be a Big Deal. (Relatively huge, nonvolatile, fast memory with a direct, fast pipe to the CPU--no SATA interrupt bottleneck bus.) 2) Pricing, as well as performance, will have to be somewhere between Main Memory DIMM and SSD/Flash Memory. 3) "Full Support" means both the Processor and the OS need to be aware, and to be fully realized, the application stack needs to exploit. This will not happen in the next two years, but over time, it will offer 10x to 100x or even more, performance boost. To me, Storage Class Memory architecture (HW, OS, and Application) is the next revolution in computing. So far, IMHO, Optane is an evolution, and nothing more than a nonvolatile DIMM. Huh?
Great summary of 2018! One correction is the perception that RDMA-based NVMe-oF requires lossless networks. While this may be true for some solutions, at least two (Attala, Kazan Networks) work fine on lossy networks, with the caveat that no protocol will work well on poorly designed or vastly under-provisioned networks. NVMe/TCP holds a lot of promise, but those fearful of lossless networks need not wait for that.
So interesting to see all the new arrays come up chasing a disappearing market. But people love those beautiful bezels! They are so cute.
Server-based storage is growing at 5X array-based storage over the past 3 years, so who needs that stuff unless you are looking to make your racks and aisles a designer playpen. Which side do you want to be on, the new and interesting way of tackling storage or are you really an interior designer on the inside? Get some stones and move on from the 20th Century already.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
