Being in the privileged position of being an editor allows me to get some interesting insights from some big players from a number of adjacent but often siloed technology areas.
For example, I have been playing a lot with unconventional storage devices lately, while witnessing some incredible work by HP with Moonshot, keeping a keen eye on what ARM and its partners are doing in the server market and what could be a glimpse at the future of personal computing.
From what I gather, the global storage market is morphing into a bipolar one with the traditional market stuck in between.
On one side, BYOD and cloud making remote storage strategies more prevalent than an on-site/in-hand, always available option. This explains why the entry level storage capacity for smartphones and tablets hasn’t surged beyond the 16GB point and why you still find laptops and desktops with 320GB hard drives, more than 10 years after that capacity was first introduced.
If you need extra storage, you’re encouraged to use a microSD card slot, a cloud storage service or simply substitute your files for an online service (e.g. Netflix, Youtube, Spotify etc). In other words, the industry is working hard to make sure that customers don’t carry more storage capacity around than they need to, permanently. You might use an SD card, a USB flash drive or an external hard drive but these were always meant to be temporary means of shuffling data around, not permanent ones.
On the other side, the advent of mega datacenters means that many are looking at ways to simplify the way data is being processed and stored. Solutions like HP’s Moonshot and AMD’s Seamicro show where the market may be going; building things that are getting bigger with components that are getting smaller (or doing more per unit volume/mass).
Small is something the tech industry knows very well. After all, where would we be without VLSI, SoC, ICs and decreasing geometries? But yet, if you look at servers over the past two decades, they haven’t changed significantly, design-wise, despite a changing landscape where demand for remote storage – with little processing power - has been growing fast, very fast.
Now have a look at HP’s Moonshot, it is essentially a smaller blade server where the biggest unit is still the motherboard.
So here’s a curveball for the industry. Should hard disk drives become smarter? Should the solid state or hard drive subsume the rest of the server (bar the power supply unit and any external connectors of course)? Should servers become even more integrated? Should the compute and the storage functions become one?
Given the propensity of the tech industry to integrate as many functionalities on die (or on chip), I wouldn’t be surprised if sooner, rather than later, system-on-drive solutions become commonplace. The key lies in how far silicon designers like LSI, Atmel or Marvell and hard drive makers like Toshiba, Seagate and Western Digital are ready to go in dabbling with the hard drive controller board. Radical changes could include
/having dedicated compute resources for different vertical tasks (encryption, deduplication, compression etc) and for traditional server workloads (database, web server, terminal, VDI, warehousing).
/revisiting the drive enclosure and the drive chassis to enhance the management of a highly integrated storage unit.
/analyzing the data flow from the compute resources with a view of reducing latency. This might include overhauling cache and buffer memory and integrating system memory somewhere.
/reviving the old 5.25in hard disk drive form factor, a move that would allow storage capacities to increase significantly as well.
/looking at alternative ways of powering a system-on-drive (Power-Over-Ethernet, Power Line communication) and deliver data simultaneously.
/rethinking the role and the relevance of physical interfaces.
Perhaps the enterprise market should look at what happened to consumer devices where extreme integration and miniaturization has been the norm, a process that has driven down prices while accelerating the growth of the user base. The smartphone is perhaps the most telling example of all; the “black hole of technology” is rapidly making a number of other devices and platforms obsolete altogether; (paper) books, personal digital assistants, Rolodex, voice recorders, movie, radio and music player, digital camera et al.
It is unlikely that these products would be available to anyone other than big OEMs or mega data center customers who have the resources and the drive (pun intended) to come up with out-of-the-box ideas to push forward the storage market. However, as always, expect some of the ideas to trickle down to the consumer market in due time.