[1] https://vimeo.com/230142234 (slides: https://www.slideshare.net/bcantrill/platform-as-reflection-...)
Of course I'm not suggesting that you forsake your values. I'm sure anyone who has followed you for any amount of time knows that will never happen. There is no stronger advocate for the principles upon which your work is based, and your fervor and commitment is extremely admirable.
My suggestion is merely that it's perhaps time to acknowledge that the market is not going to adopt illumos in a major way. An outflow of that is considering the possibility that the considerable engineering resources at your command may be better deployed in attempting to instill and carry forth the same crucial, badly-needed principles and values into projects that have more realistic prospects, to the extent possible, rather than remaining 1000% committed to a lost cause, where the principles and values will flare out in obscurity. "If a tree falls in the forest..."
illumos is an immense contribution to our collective engineering heritage and I entertain no illusion that it should be totally iced. Personally I would consider it a glorious resurgence to see it more widely adopted, and to see it become the basis of a new generation of systems. SmartOS is simply a joy compared to the train wreck that passes for systems management in today's mainstream, and I sincerely hope that the wheels turn to make it a more viable option. The ongoing open-source contributions of the faithful volunteer keepers of the flame will be invaluable in leaving that possibility open.
While there is no doubt that illumos's open-source heritage ensures it will live on, we must be realistic and admit that for all practical intents and purposes, the sun has set on the project's general commercial viability (at least for the time being -- the thing about sunsets is, they're followed by sunrises).
The question I pose is whether your advocacy may be more effective in bringing the values and contributions of yourself and your team in force to more active venues, than by continuing to toil in the last fleeting fragments of the twilight of Solaris's commercial viability.
I personally know many young engineers who admire your talks but have no exposure to the actual process of working with you directly. What would change if they got that? Would there be a Cantrill Youth? What's the likelihood of passing on that flame within illumos v. expanding your horizons to other groups?
Sometimes a reference implementation is intended to be just that, and nothing more—a reference, for other implementors of other systems to crib ideas from, especially as pertains to how the smaller ideas cohere architecturally in practice (which is hard to communicate by just writing papers about the individual ideas, or building them as standalone libraries, or building them into other systems where they must interoperate with components that don’t cohere into the desired architecture.)
Or, to put that another way: sometimes the best textbook is the codebase of a good implementation. And some people are more interested in writing those textbooks, than in personally getting their ideas into production systems. Especially because it’s far easier to refine the ideas in the context of such a coherent presentation of them.
Someone else, who reads such a “textbook” codebase, can put the ideas to work. But the ideas themselves will be better-developed for having had time to stew in the same pot as other novel ideas. This serves the people who put the ideas into practice as well.
I'm sure they're fine with it being Their Own OS that's only actually used for the Samsung Internal Cloud Thingy, while the source is public for anyone to take so it won't disappear from history.
You're asking Rembrandt to stop painting realistic paintings because Picasso nowadays generally passes for art. What you don't consider is that there is nothing stopping Rembrandt from continuing to paint his realistic paintings for his own sake and pleasure.
This is by far the best explanation I have heard. Suddenly Joyent makes sense to me after all these years. They have been continuing to do what Sun did.
In a very similar way to how Sun followed the fatal strategy of simultaneously trying to make money selling proprietary Unix hardware, while also making their software open source and porting it to more popular architectures, while simultaneously developing Java as a portable programming language intending to erase all differentiation in underlying architecture/OS differences …
Yes: Joyent is (was) a hosting company trying to compete with AWS, claiming technical superiority in the software stack. Yes they also make an OpenStack competitor. And had their own programming language runtime, node.js, which seemed to be an odd implementation choice for their classic old-school unix tech. All of the above.
Sun would have done better to have made SPARC hardware more accessible than to have open sourced Solaris. Hindsight is 20/20, but now that we've reached the limits of single-threaded performance Sun's emphasis on multi-threading and specialized ISA extensions would have made both Solaris and SPARC competitive today.
Getting there would have required cannibalizing their enterprise income, though, and that's difficult if not impossible for any company to do. They made the gamble on Solaris because it was less risky--major enterprises were always going to stick to Sun's Solaris--but low-end SPARC hardware absolutely would have hurt their bottom line.
That this came from a company essentially selling mini computers.
In fact, we are trying to do multi-tenant all over again. To my mind, this cycle of containerization is still trying to fulfill all the promises we were given in the early 90's about memory protected, pre-emptive multitasking OSes that never came true. Which itself wasn't the first time (it's all copying mainframe ideas, badly). Given all of the problems coming out about speculative execution and data breaches, I'm already curious what the next set of promises will look like.
I man not really, Joyent is a hosting company (cloud & virtualisation). That's built upon an Illumos distribution (SmartOS) but SmartOS is not their business.
SmartOS may've been indirectly implicated in revenue, but let's be real, illumos and Joyent have always been tied at the hip. Joyent was never going to become a Linux-centric alternative public cloud like AWS, GCloud, etc.
If there had been less ideological fervor and with Samsung's backing, maybe they could've offered more serious competition to the incumbents, but the Solaris style oozes out of Joyent's every pore. It's fundamental to their way of thinking. That purity and commitment is certainly admirable, but it's simply not commercially viable anymore.
Joyent is proof that large scale cloud can be done with SmartOS.
The death knell really being the abandonment of illumos-gate as the authoritative ZFS upstream. If old-school diehards like Matt Ahrens are throwing up their hands and saying that the momentum in ZFS is in ZoL, the battle is over IMO.
node.js was always a diversion IMO. I think it was a play to maintain relevance, something Joyent has always acknowledged a need to do, but I don't think they've ever had the manpower or the buy-in to go full throttle.
I don't know if this is the end of Joyent as such or not, but IMO it's finally time for Cantrill et al to turn in. They've undoubtedly fought the good fight, but illumos is a farce at this point. Better to accept that their old-school big Unix development model is a thing of the past and put those efforts to something more practically useful, like FreeBSD. At this point, they're torturing the legacy of Solaris more than maintaining it.