I am neither a software engineer, hardware engineer, nor electrical engineer, but I did stay at a Holiday Inn Express last night (obscure, questionably humorous ad reference). Technically I am a Data Wrangler, an Oracle DBA (Database Administrator), a SysAdmin (Systems Adminstrator), a troubleshooter, yada, yada, yada. I have a keen interest in all things technological. I am a geek (if that is a positive accolade) and was a nerd (a negative accolade) as a kid. I have some cred.  

I present for your consideration that monolithic memory is the holy grail of computing and within our grasp in the next few years. This is good news for consumers, not so much for old guard industries.

Consider the modern computer. By computer, I include desktops, laptops, servers, tablets, phablets, smart phones, whatever … running any operating system including but not limited to Windows, Linux, Unix, MacOS, iOS, WatchOS, Android, DOS … you get the idea. Regardless of manufacturer, these systems are all remarkably similar. They each have one or more CPUs (Central Processing Units), each CPU having one or more levels of dedicated ultra-high-speed memory called cache. 

Next, they each have a shared block of high-speed RAM (Random Access Memory) which is dynamic (hence DRAM). DRAM is fast. But that speed comes at a price. All data is lost when power is turned off. Recall the time you forgot to save that epic document or spreadsheet and the power went out? Yeah, that drawback.

Finally there is storage, usually in the form of a hard disk drive (HDD), although more and more computers use some form of solid state or flash storage (SSD for Solid State Drive). Mobile devices make heavy use of flash storage. Storage is persistent, but slow. Historical forms of storage include floppy disks, magnetic tape, and even paper punch tape and punch cards. Slow, but persistent.

The term memory is used contextually to describe each of these “data buckets”. The statement “I have 16 gig of memory”, is ambiguous without context. Do you have 16 GB (gigabytes) of RAM in your laptop? Or 16 GB of storage on your iPad? The former is a lot unless you are a gamer or scientist. The later is woefully small especially if you want to store a video or audio collection. RAM is currently supplied in tens of GB and usually in powers of 2: 1, 2, 4, 8, 16, 32, 64, 128 GB. Storage on the other hand is now commonly hundreds and thousands of GB.

Modern computer systems have evolved the subsystems necessary to route data between the CPU, cache, dynamic RAM, and persistent storage. These subsystems are comprised of both hardware (controllers, data busses, I/O channels, etc.) and software (drivers, modules, packages, etc). Data flow is a well choreographed dance between low, medium, and high-speed subsystems and pipelines. A true monolithic memory system would eliminate the need for all of this. No more need for swap or page files. No more paging of memory out to disk. No more “saving” work out to disk. No more disk.

In the previous post, I addressed claims by Nantero that their carbon-nanotube-based NRAM offers the tantalizing possibility of lower power, higher data density, faster response, and lower cost than all other types of conventional memory. NRAM has the potential to provide the basis for true monolithic memory. But it won’t happen overnight. Even assuming that NRAM (or a competing technology) is up to the task, no existing operating system or hardware platform is up to the task. A complete redesign of memory management of both the hardware and operating system would be required. 

As a consumer, this is great news. All consumer computer devices will become like smart phones and tablets from the user’s point of view. Always on, instant “save”, super fast. Except that now the amount of storage will be many times greater. Power consumption will be primarily a factor of display efficiency. Speed and power will be better in every way. Prices will fall as capability increases. The consumer wins all the way around.

Not everyone will be a winner. Old school RAM, flash, and hard drive manufactures will have an uphill fight to remain relevant. Their investors will suffer as the share value of these companies fall. Mergers and acquisitions will contract the industry like a collapsing blackhole. History repeats. How many steam locomotive or buggy whip manufactures can you name?