It's been an interesting few decades since Oracle and Sun Microsystems argued that we should shift computing from the desktop to servers and first created somewhat-viable thin clients - when I say "interesting," I mean the solutions. From the standpoint of the user and in terms of cost, generally stunk.
However, just like it took a few decades for tablets to reach the iPad stage, when they were accepted with the right technology, backing and execution, the idea of the thin client was actually a good one. We were just waiting for the right offering.
Well, it just emerged, driven by VMware and Nvidia. It blends PC and server architectures with some new technologies to create something amazing. Let's talk about what was wrong with the early efforts and why this VMware/Nvidia move, Horizon View 5.3, has the right stuff.
Thin Clients Never Lived Up to Their Promise
The promise of thin client architecture was compelling: The performance of a PC with the security of a mainframe and the reliability of an appliance. The initial problem was that the two guys making the big push, Scott McNealy of Sun and Larry Ellison of Oracle, though they clearly understood the security and reliability aspects of the solution, couldn't spell PC. They also disagreed on cost.
Meeting a bunch of European CIOs back in the day, the Sun executives in the room (who were supposed to be silent observers) suddenly jumped in with outspoken criticism of Microsoft. The CIO tore them a new rear one, saying that, as bad as Microsoft was, a traditional Windows solution was massively better than what Sun proposed, largely because it was both cheaper and better aligned with what users wanted. Sun did deploy its Sun Ray solution internally, which I believe contributed materially to Sun's collapse. Oracle, meanwhile, didn't deploy its solution broadly and ended up buying what was left of Sun.
Over time, two camps emerged: One led by Clear Cube, which makes blade PCs, and another eventually led by Dell and HP, who use a server backend to provide their version of a thin client increasingly wrapped with Citrix XenDesktop. The former wouldn't scale but provided PC-like performance and cost; the latter was less expensive in hardware and scaled well, but it was more expensive in software and lagged in performance. Oh, and neither solution covered mobile employees well.
What was needed wasn't a PC-based solution packaged like a server or a server-based solution packaged like a hosted PC but, rather, an attractively priced new architecture. My contention: The entire solution needed to be rethought - not by server guys, but by PC guys, because it had to be as good as or better than PCs to be successful. Few seemed eager to buy a compromised solution that never threatened the PC market model.