Open Stack

On 02/11/2016 05:46 PM, Jay Pipes wrote:
>> *From:*Boris Renski [mailto:brenski at mirantis.com]
>> *Sent:* Thursday, February 11, 2016 3:05 PM
>> *To:* AlanClark
>> *Cc:* <foundation at lists.openstack.org>
>> *Subject:* Re: [OpenStack Foundation] Updating the OpenStack Mission
>> Statement
>>
>> Does anybody share a concern that the current mission statement is a bit
>> too generic and broad?
>>
>> Once this mission statement is published, I can see headlines "OpenStack
>> will fail because it aims to be all things to all people. Updated
>> mission statement confirms it's inevitable doom."
>
> Heh, this already has happened, without any change in mission statement!
>
> http://cloudarchitectmusings.com/2016/02/08/between-a-rock-and-a-hard-place-will-openstack-become-niche/
>
>
> For those of us who, years ago, warned about the dangers of "enterprise"
> workloads -- i.e. pets -- not being a good fit for OpenStack, and how
> the future was cattle workloads and cloud-native apps, this article
> kinda hit a nerve.

I would argue the opposite - that chasing so-called "cloud native" 
workloads is and always has been a mistake for us, and that we've never 
been good at that - nor are any VM-based clouds. Kubernetes and Swarm 
are good at them. OpenStack is not.

I don't want to be an Open Source AWS Alternative. I also don't want to 
be an Open Source VMWare Alternative. Being an Open Source Alternative 
has literally never worked. Being the thing that you are - even when the 
conventional wisdom of the time says it's the wrong thing to want to be 
... that's the thing that has been successful.

OpenStack provides excellent API-driven ops for people who want things 
that behave like actual computers. It is now and always will be bad at 
dealing with workloads where people want sub-second start times and 
where the average lifespan of a workload is 5 seconds. If you actually 
want "cloud native", we're the wrong place to look.

We didn't actually build a thing to do ephemeral workloads - we made a 
thing that's actually pretty amazing at running persistent computers. I 
really wish we'd all stop trying to be something we are not and never 
will be and instead revel in the thing we actually are.

> But then again, I keep hearing that these enterprises are paying all the
> salaries... so... who knows?

Maybe that shows that we have a real value - even if it's not the value 
that people thought it was going to be six years ago. I, for one, very 
much value that in OpenStack clouds (and not AWS, Google or Azure) I can 
get VMs in public clouds with actual routed IPs. That's amazing.

I value that I have had a VM in an OpenStack cloud at Rackspace running 
with no downtime for a few YEARS - and that I have a VM running in an 
OpenStack cloud at Vexxhost has not gone down since I spun it up 194 
days ago. I value that I interact with both in the same way.

I value that there are Public OpenStack clouds today based in Canada, 
Japan, China, Sweden, Italy, The UK, The Czech Republic and oh also the 
US. I value that because I do not think that the US is the center of the 
world and I do not think that people who do not live in the US should 
have their hopes and desires gated by the profit motives of companies in 
the US. I value that if someone in Uruguay wants Cloud Computing, they 
do not have to hope and pray that a product manager in Seattle, 
Washington runs the cost/benefit and decides in their infinite wisdom 
from on high to install a data center there - and then charge the locals 
prices based on the economy of the US West Coast. I value that we are 
succeeding at this today.

I value that all of this is possible even in the face of people who have 
theories of how things *should* be done and who tell people that what 
they desire is wrong for them to want.

I am sick and tired of people who tell me that I should instead want a 
system that is complete epic overkill for most things.

I'm tired of the "conventional wisdom" being that I should essentially 
ALWAYS build an application that could scale to Facebook size - even 
though almost no applications built have a need for sharding, sub-second 
failover or even running on more than one computer.

I want to make a thing that the Powers That Be are refusing to make.

I want to make a thing that is what people WANT but are being told they 
should be ashamed for wanting.

I want to make a thing that I WANT to use. And I'd like to point out 
that I'm excellent at both Dev and at Ops.

So here's my mission statement:

To produce a single global Open Source Cloud that is federated across 
Public and Private installations that provides API-driven access to 
resources that look and behave like Internet connected Computers and 
that re-decentralizes the Internet from the miasma of walled gardens 
inflicted upon the world by the Oligarchy of US Billionaires.

I'm sure we won't select it as our Mission - but it's the mission I've 
been working on for the last six years - and it's the mission I'll be 
working on for as long as I'm here.

Monty