~30 of Hypervisor: From Magic to Commodity

I was cleaning out the attic a few days ago when I stumbled upon an old piece of paper that stopped me in my tracks. It turns out, today is exactly 20 years since I grabbed my first VMware certification. I’d already been working with their products like Workstation and the early versions of ESX for years before that exam, but the paper makes it official.

Seeing that certificate triggered a specific memory from the ESX 2.0 days. I had just finished testing the new version and headed out to a consulting gig. The job was actually for some Linux work, I’ve completely forgotten what the specific task was, but I was so hyped on ESX that I just had to pitch it to them.

I’ll never forget the client’s response. He looked at me, held his thumb and index finger a tiny bit apart, and said, "We already run it, this is how much we paid for the license." Then he threw his arms wide open and said, "And THIS is what we got."

That was the moment it clicked. I already knew the tech was brilliant, but that reaction proved it was going to sell.

Is the Hypervisor "Done"?

Looking at the landscape today, the hypervisor has basically become a commodity. It makes you wonder: are we ever going to see those "mic drop" features again? I miss the days when things like vMotion, HA, DRS, or VDS were brand new and in some ways mind-blowing.

I remember when VMware added the ability to record and replay CPU instructions in Workstation. I thought to myself, "I bet they bring this to ESX as some kind of enhanced HA feature." And they did. Technically, Fault Tolerance (FT) is fascinating engineering, even if it never really found a major use case in the real world.

But what’s next?

Storage and Networking (the rest of the SDDC) are guaranteed to keep evolving with big new features. But the hypervisor? I’m skeptical. Maybe one day we’ll be able to vMotion a single process from one VM to another? (Is there actually a use case for this, though?) Or is there something else out there?

Otherwise, it feels like the hypervisor development has stagnated. It’s "finished." The future of on-prem development seems to be entirely about the Private Cloud layer on top, rather than the engine underneath, although network and storage parts will probably still evolve —and that’s a good thing.

Coming Up Next?

Finding this paper in the attic didn't just bring back great memories; it got me thinking about the actual value of collecting these badges.

If I write a follow-up to this, I’m going to dig into whether getting certified is actually worth it these days. I will try to utilize the experience I have gathered over the years as a VMware instructor, teaching classes in different countries where people have very different views on the matter.