[OmniOS-discuss] The Future of OmniOS
robert at omniti.com
Fri Apr 21 14:07:20 UTC 2017
Five years ago, when we first launched OmniOS, we did it out of a
direct need to push forward the OpenSolaris ecosystem that we had
built into the core of several parts of our business. At the time, the
illumos community was still rather new and taking direct control of
our path forward was a solid next step; we had already built many of
the pieces in-house that we needed to produce a complete operating
system distribution, and our experiences with open sourcing software
we worked on had been generally very good.
While we didn't know quite what the reaction would be, there were two
things internally that guided us as long term factors in our decision.
First, as we have done for other open source software, we thought it
made sense to offer commercial support for OmniOS, but there was no
desire to "pivot" OmniTI to be an operating system vendor. We like the
world of building and running high-scale software and infrastructure
and that's where we wanted to stay. Hand in hand with that was the
second idea, that while we felt it was important for us to take the
first initial steps, in the long term we really would prefer that
OmniOS become an open source project maintained by its community
rather than remain as the open source product of a single commercial
entity (think Debian vs Red Hat, if that helps).
Five years later, we are proud to see that this software has been
accepted by a wide group of companies and end users, and we think this
has been a boon for the illumos community, who are the shoulders we
build upon. When you see companies from all sectors and industry, both
small and some orders of magnitude larger, using the technology you
put forward to build even further; well, it's great to have an impact.
However, even with the success we have had, there is one area we have
failed to make progress on, which is the goal of making OmniOS
community operated. There are many factors why this hasn't happened,
but ultimately in five years of both ups and downs within OmniTI, I am
left to conclude that if we are ever to change the nature of OmniOS,
we need to take a radical approach.
Therefore, going forward, while some of our staff may continue
contributing, OmniTI will be suspending active development of OmniOS.
Our next release, currently in beta, will become the final release
from OmniTI. We are currently going through steps to remove any build
dependencies on OmniTI or its infrastructure, and we've made some
steps towards determining what potential resources we currently
control which could be turned over to an open source community should
one emerge; for example, we can continue running OmniOS mailing lists
from OmniTI, but would eventually like to see those transitioned to
something operated by the community itself.
To be clear, our goal is not to abandon OmniOS, but to divest OmniTI
from the open source project in order to spur others to participate
more. We still run quite a bit of infrastructure on OmniOS and expect
to continue contributing, but the current model does not work for
OmniTI nor do we believe it is healthy for the OmniOS community as a
whole. Could this mean the end of OmniOS? We can't guarantee it won't.
For that matter, recent user data shows that a majority of the
community still uses OmniOS primarily as a storage solution, not a
platform for high-scale web computing (which was our original intent),
so even if a community does form, it could move the project in a
direction that doesn't align with our needs. If that happens, we feel
comfortable knowing there are several other strong illumos based
options available. In the end, while this rip-the-band-aid-off
approach is not without risk, it is one we feel is necessary.
We hope that most folks will respond to this not with fear but with
the understanding that there is now an opportunity to build a broader,
stronger community, and we look forward to working with others to make
that a reality.
More information about the OmniOS-discuss