Here Brendan Gregg who was a kernel and performance Solaris engineer who worked at Sun Microsystems and later at Oracle Corporation following its acquisition by Sun is recommending to all former Solaris users to switch to Linux
http://www.brendangregg.com/blog/2017-09-05/solaris-to-linux-2017.html
I assume that this blog post is more job related then technology related, its a lot easier to find Linux job then to find FreeBSD job, and its a lob easier to find FreeBSD job then Illumos job.
He also mentions BSD's several times but his 'focus' on Linux technologies are superficial at most.
He mentions Docker with Zones in mind but he does not mentions that Docker has ZERO separation without SELinux (or AppArmor), which are PITA to configure and use. Same with networking, Solaris/Illumos has great resource control via projects, along with creating a separate SMF service with two processes with separate CPU pool and different CPU scheduling for that pool ... but Linux only has network namespaces and cgroups ... its like comparing SR-71 Blackbird (Solaris) to Cessna 172 (Linux). Because something is popular it does not make it better. Going that route, lets eat sh!t, millions of flies can not be wrong ...
So this blog post is 'job' biased (which is not bad but You need to keep that in mind while reading it).
I use FreeBSD since 2005 (FreeBSD 5.x) and I use it till this very day and will be using it in the future, but is my job FreeBSD focused? Unfortunately no. Can I use FreeBSD at work? Sometimes when its appreciated and better suited then Linux or better suited then IBM AIX etc., for example for storage servers with ZFS and its features, sometimes for Jails, sometimes for other features ...
IIRC
vermaden was a serious OpenSolaris user so he probably can shine more lite to the future of Illumos kernel. IMHO as somebody who grow up using Solaris (my third UNIX after Tru64 and short affair with Irix) starting with Solaris nowadays is an example of necrophilia. Illumos kernel IMHO has no future.
Illumos future is not great, but still better then 'official' Oracle Solaris after all these RIF layoffs of almost entire Solaris and SPARC teams. I would compare that to numbers, if Linux ecosystem size is 100, then FreeBSD ecosystem size is 10 and Illumos ecosystem is 1, (which is very generalized).
Do Illumos has future? Sure. SAMSUNG has bought SmartOS 'cloud os' (Joyent) which is great in what is used for and I doubt it will 'die' anytime soon, many Stock Markets use SmartOS for their services as its mentioned in this video:
https://containersummit.io/events/sf-2015/videos/wolf-of-what-containers-on-wall-street
Same for Delphix which uses Illumos in its database appliances.
Of course one can always decide to purchase
NexentaStor if they were previously invested heavily in Solaris but in my experience most companies which have no in house expertise to run vanilla FreeBSD these days will prefer
TrueNAS in spite of Corral fiasco and PC-BSD to TrueOS monkey business which affected only IXSystems unpaid customers.
I would omit NexentaStor personally, FreeNAS is far better solution here (and free for all purposes).
KVM is a Linux thing and running it on SmartOS sounds dumb to me.
If something is dumb, but works, its not dumb. Why write your own hypervisor while You can port one? FreeBSD's Bhyve also has been ported to Illumos recently. 'Funny' thing about KVM on SmartOS/Illumos is that its Intel only, does not support AMD.
Its also one of the most important things that are missing on Oracle Solaris, X86 virtualization. In Sun times there were xVM project which integrated XEN and Solaris and worked pretty well, but as Oracle took over Sun they killed it and created Oracle VM for X86 which is base on Red Hat Linux clone - Oracle Linux and that still leaves Solaris without ANY X86 virtualization possibilities in 2018, which IS dumb.
The very fact that SmartOS uses NetBSD pkgsrc as the official package management should be a big red flag.
Probably the same thing as with KVM, You can do it yourself from the ground up and 'waste' more time or get something that already works and polish it a little, for the purposes of SmartOS (almost non changable virtualization/cloud host) this seems to be good choice, it would be other case if they would like to be 'general' distribution like OpenIndiana or like FreeBSD.
Personally when it comes to virtualization I am heavily vested in Xen and my choice for Dom0 is Alpine Linux.
You should post some HOWTO on FreeBSD formus, I would be very interested to see how its setup, as Alpine Linux seems to be sh!t free Linux, or at least from most of the typical Linux sh!t like systemd or glibc.
OmniOS is dead. As much as I liked Jason Dixon as OpenBSD contributor his OmniIT Labs never produced a useful working open source tool (I hope paid customers did better).
OmniOS has just been 'phased' to community with introduction of OmniOSce (Community Edition) and they already started to provide 'paid' version (which is good progress in they journey). Will tehy survive? Maybe. Its more about Illumos ecosystem then about OmniOSce.
For example, If You want 'desktop' Illumos, then You get OpenIndiana, if You want cloud, then You get SmartOS, if You want server Illumis, then You get OmniOSce, there is 'almost' nothing more ...
I would prefer that Illumos would go the FreeBSD way with general purpose core and maybe some variants, but the model of central 'general' core of Illumos and 'external' distributions seems to not work that well.
I don't see anything on that page that deserves any further comment. By the way Ubuntu 16.04 "ships with ZFS" and due to the fact that most OpenZFS contributors use Linux at work I would not be surprised that ZFS becomes first class citizen on Linux now that Red Hat removed BTRFS vaporware from their OS.
But can You install Ubuntu on ZFS root? Nope. So its the same 'ZFS on Linux' bullsh!t as on every other Linux distribution. You can not use ZFS Boot Environments with Ubuntu (or any other Linux). ZFS on Linux is still sh!t.
Obvious problem is that Oracle owns ZFS but for now all of us who are using it chose to ignore that fact (myself included).
Oracle owns their own ZFS but Oracle DOES NOT OWN OpenZFS, so Oracle is not a problem here.
BTW, OpenZFS 'joint project', which takes Illumos, FreeBSD, Linux and Mac OS X 'distributions' into OpenZFS still keeps the 'core' source in Illumos ...