I have been thinking for a long time that tried and true operating system architecture that we have relied upon since the earliest days of commercial computing might be holding us back. What I’m getting at is the built in assumption that most OS designs have that the hardware consists of a processor (CPU), volatile memory (RAM), and offline storage (Disk Drive). With the advent of solid state drives in capacities that are competitive with traditional technologies (e.g.: Intel thinks big with solid-state drives | Nanotech: The Circuits Blog – CNET News ) and with processors that include larger and larger on-board cache, although I’m no electronics engineer, it seems plausible that some kind of convergence could happen if we can break the assumptions that OS designs are built upon.
A lot of what a computer operating system is dedicated to doing is managing the loading and unloading of programs and data into and out of memory to prepare for execution and manipulation by the processor.
What if programs could be executed in place?
What if data could be manipulated in place?
How much computing power could be redirected to serving the user vs. serving the hardware?
Could we have instant resume, instance suspend computers (and eliminate the concept of "on" and "off")?
Could we have a computer that didn’t crash (in the sense of a hard disk crash)?
Could we have a computer that operated for months on a single battery charge?
Could we have a computer with the computing power of a laptop, but that folded up to fit into your wallet?
It’s a bit of a chicken and egg problem, I suppose. I think, based on how things worked out last time around that we need to get creative on the hardware architecture side and make that drive the operating system. What if we rewound history back to the point the breakthrough of the loadable programs was made and solved that problem with today’s technology, in a different way. What if the operating system’s job was to constantly reconfigure a programmable logic device of a few hundred gigabytes? I think things might be different than the current "state of the art" in consumer computing if we could affect a shift to adopt reconfigurable computing in the mainstream.