Tomorrow (Friday 15th Dec) at 11am in the ‘Progress’ meeting room at the Federation in Manchester a small group of us are discussing the potential for a cooperative hosting project as part of the broader CoTech network.
We’ll publish notes etc from the meeting on the Nextcloud instance in the folder ‘CoTech’ > ‘Cooperative hosting’ and encourage anyone with an interest to get involved.
I think GreenNet would be interested in getting involved with this (though
of course we’re still not formally CoTech members, hopefully that will
change soon)
Thanks to all for taking the time to get involved in the discussion, which I thought was positive and has great potential. Lots still to tease out and clarify, but I got a sense that do have a real opportunity here that is worthy of further investigation.
Please take the time to add to and edit these notes, and/or begin work on associated documents in this folder.
I wasn’t present today but have a keen interest in this myself, in particular as a viable alternative to AWS. That is a huge task though, as modern application development practices (12 factor and the like) demand SaaS and PaaS, not just IaaS.
From what I see above you have a choice between a commercially viable AWS replacement that aspires to be community directed (but professionally executed) or a more values driven proposition which would be unlikely to compete on price or offering with AWS or a near competitor of it.
I heavily favour the former, but that’s just me. What was the feeling in the room leaning towards?
We know that launching a cloud service - probably using OpenStack - will be hard. So we want to do other things first - shared web, virtual servers, domains. We want to get to the point where we have these products being served from one of the co-operatively owned/managed datacentres that we can access: so that the whole stack will be co-operative. Then we will look at a cloud service.
I’ve spent quite a long time looking at OpenNebula, which is much simpler to run than OpenStack. I was considering it as a fairly good initial offering for a co-op like the one that is being discussed, but stalled largely due to time and hitting a wall with my knowledge of networking.
In any case, OpenNebula offers IaaS and is very extensible. I’ve actually been working on a local development environment for it so I can work out user provisioning. It isn’t impossibly difficult although the API isn’t as nice as OS or AWS.
It is actually doable, I think. It’d even be reasonable simple to do sensible CI.
I’m based in York and now work full time in Sheffield: I gather some people here are based in Sheffield, maybe we could chat over a drink?
Since we (Webarchitects) do shared web, virtual servers and domains already the next step for us would be to look at how to deploy something like OpenStack or OpenNebula (looks interesting, hadn’t come across it before) — creating a new co-op consortium to offer the same services as we offer currently doesn’t really make much sense to me… or am I missing something?
Offering shared web, virtual servers and domains, and more, for example OpenShift, provisioned using GitLab CI on top of OpenStack or OpenNebula is the direction I think we would be interested in going in (note that GitLab appears to be heading down the Kubernetes road)…
Currently we don’t have CI setup at git.coop but I’m going to see if I can deploy that over the Christmas holiday.
In terms of the sense (or otherwise) of the initial service offering, my understanding from our discussion yesterday was that we are taking an evolutionary approach - starting where we are currently, establishing the basics, exploring the immediate opportunities and learning what will work by way of the likes of AWS alternatives, etc., so that we can expand into new service offerings over time.