On 2015-05-17 at 09:26 -0500, Matt Lawrence wrote:
> I've heard from multiple people, who I respect greatly, that OpenStack is
> going to die off, but I just don't see it. The modern IT infrastructure is
> heavily populated with products and technologies that were a really bad idea
> in their original release. So, while I think OpenStack has some serious
> problems, I think it is going to be around for quite a while.
>
> I would like to hear what a number of other folks who I respect think.
Disclaimer: I work for a private/hybrid cloud vendor (and we're hiring),
but we're container providers, so still sit on top of IaaS clouds.
I think that the opinions expressed are mostly optimism that "there has
to be something better coming". For this layer, in this price-point,
I'm skeptical that there is. I think that OpenStack will change in ways
so that you'll end up with something still called OpenStack, but with
the worst pain-points replaced. So my skepticism is actually optimism,
because I think that what we call "OpenStack" stands a chance of
improving. It's already at the point where it solves more problems than
it causes. As a cynic, that's high praise from me.
If you've been around a while, you might remember DCE, the Distributed
Computing Environment, which was a 1990s multi-vendor mish-mash of
pieces thrown together; different pieces from different vendors, of
varying quality. Some parts were excellent, some not so much. Trying
to program to DCE was a nightmare, because each different component had
completely different conventions, starting at the "who allocates memory
and who is responsible for freeing it" and moving up from there. Some
parts of DCE live on, mostly the parts which already existed and where
DCE "just" made a fork (Kerberos, AFS, etc).
In various ways, OpenStack reminds me of DCE, but I think that it has a
much higher chance of remaining relevant.
So far, OpenStack has had some excellent components and some horrible
ones. The authorization provider, Keystone, which should be one of the
most secure parts because grants access for everything else, has a
_horrible_ CVE history.
A key point is that where DCE was one fixed stack which various OS
vendors will or won't get around to implementing the parts of, badly,
with different commercial interests pressing them away, with OpenStack
the OS vendors play second fiddle unless they can make themselves stay
relevant by offering nice pre-canned setups. With OpenStack, the
company putting together the components gets to decide which to include,
they could opt to replace something used by others, and that company, as
the technology chooser, has a vested financial interest in making their
product successful. No DCE undermining of compatibility ("you don't
really want your code able to run on our competitors' products, these
are the only bits you should care about").
So there's market competition between OpenStack vendors. Those who make
good choices will survive, those who make bad choices won't, while we as
technical folk can only hope that what constitutes a "good" business
choice is also a good technical choice. Optimistically though, that
competition between players each trying to make their product attract
customers should help things keep moving forward.
--
https://www.apcera.com/
_______________________________________________
Discuss mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
http://lopsa.org/
