OpenStack at 10 years old: A failure on its own terms, a success in its own niche

When OpenStack announced itself on July 19, 2010, it gave itself a mission “to produce the ubiquitous OpenSource Cloud Computing platform that will meet the needs of public and private clouds regardless of size, by being simple to implement and massively scalable.” Measured against that mission statement, the project has …

[Anonymous Coward]

"platform that is easy to use, simple to implement"

Good luck with that - they can start avoiding the need to change the very same settings in many, many files at setup.... a better UI to monitor and manage the solution would also be welcome.

No surprise ready solutions like VMWare VIO are often used to minimize the implementation effort....

[Jusme]

Re: "platform that is easy to use, simple to implement"

Last time I looked, OpenStack was still a "science project". You get a bag of bits (very nice bits, certainly), but putting them together to make a working virtualisation environment takes a lot of time/effort/knowledge. Compare with VMware, which "just works" (though they're trying their best to break it with every new release), and you can be spinning up VMs through a nice-ish GUI within minutes of installation.

OpenStack needs to be a product, not a technology...

[Todd-InMotion]

Re: "platform that is easy to use, simple to implement"

I am a big OpenStack fan, so some bias here, but my take is we (the OpenStack community, but this is my personal opinion of course) have been “failing” in exactly what you are talking about. To encourage adoption OpenStack *must* be easier to get started with as a user. If you have to have an expert team to cover networking, hardware selection, setup automation, etc. and that is just to get your test cloud going, not actually use it!

That being said, we also have the chance every day to start succeeding in making OpenStack more accessible. For me, to spin up a new basic OpenStack cloud must be trivial so that any IT teams, small, large, 1 person, can learn it and be comfortable with it as a user before trying to be an architect.

There is big push now in the ecosystem to provide on demand openstack private clouds (in a few hours) to help solve this and a few providers are there or almost there. I think this future will jumpstart usage, here is hoping!

[dinsdale54]

Yup, difficult to use.

I did some work with OpenStack a few years ago and found it incredibly fiddly to set up and with documentation that assumed a high level of pre-existing knowledge. It reminded me of Unix in the early 90s. If you have a room full of techies to understand it all and manage the configuration then I can see how it is a good cloud platform with a lot of functionality. Otherwise, no go.

Also, a company I worked for had a product where the entire management interface was OpenStack based. The techies referred to it as Brokenstack

[IGotOut]

It will be interesting to see

If India start using it more, with there ambition to become less reliant on other nations tech.

China + India....that's a pretty big niche.

[TechnicalVault]

Work with someone who knows what they're doing

At my place we are running 2 OpenStack installations and they are quite happy little private clouds. If I had one thing to say about it, it is that we were early adopters but Openstack is a lot more mature than when we first started. There are still a few more features we would like but it has finally reached the point where it can run happily without choking everytime RabbitMQ dies.

You can go it alone and try and put something together directly from Openstack but you are probably better off working with someone who knows where all the bumps in the road are. We have just switched Openstack distributions from Redhat to StackHPC on our primary and so far so good.

[anothercynic]

Re: Work with someone who knows what they're doing

Well, StackHPC are arguably one of the best in the business... Stig knows the tech pretty much inside out and has a lot of experience in the HPC field, so... :-)

[IGnatius T Foobar !]

Nutanix is KVM and that's good enough for me.

Nutanix AHV is a KVM-based private cloud infrastructure, and it's easy to set up. If you want the goodness of KVM, and you want to avoid Amazon and Azure, Nutanix is a good way to go -- it's finding a lot of inroads there. VMware is still the big player and they're not slowing down either.

Everyone must work together against an Amazon monoculture.

[Nate Amsden]

don't forget cheap

Customers wanted the easy to use, the scalability and they didn't want to pay VMware etc for licensing. Perhaps Openstack will get a big boost from the new EU cloud project or whatever.

The move from utility computing(what I prefer, not an often used term) to (infrastructure) cloud computing comes at an astronomical relative cost, and in my opinion in most cases isn't justified. Huge cost increases come from huge increases in complexity etc.