Share with friends and colleagues on social media

FullSizeRender (1)

This is me, and this is my mother.

Another OpenStack Summit is in the books, and wow, was this one busy! 7500 people in Austin – 100x growth from the first summit held there only 5 years ago. And if you still have questions about whether OpenStack is ready for mission critical workloads, check out these keynotes by SAP, AT&T, Volkswagen, Time Warner Cable and others.

We had our largest number of SUSE sessions so far – 14 in all – and among them was an intro-level panel session with Midokura, NetApp, Redapt and Fujitsu titled “Things I wish my mother had told me about OpenStack.”

Now, my mother is pretty tech savvy. She’s 89 and has a macbook and iPhone from which she sends emails and Facebook updates. Truthfully though, my mother and I have never actually had the “OpenStack talk”, but if we had, there are a few things that I would have wanted her to set me straight on.

OpenStack Distro vs. DIY

“Mom, I’m going to deploy OpenStack at home – to manage the kids schedule, our finances, my blog, the house climate control, the vegetable garden, global warming – but I don’t know whether to do it myself or use an OpenStack distribution.”

“Son”, my mother replies with a gentle smile, “this is a great question” (she’s always been very affirmative like that). “I mean, OpenStack is open source code, so why not just grab the bits from upstream and do-it-yourself. Creating your own OpenStack distribution is like creating your own Linux distribution: You would only do it if you really, really, really needed to. In the end, for most use cases there’s no real advantage and in fact it can end up creating a lot more work. Like that time you cut your own hair. Now, starting with an OpenStack distribution, that’s where you see some real advantages, beginning with the installation framework. There are dozens of components and well over a thousand parameters that you need to configure to deploy an OpenStack cloud. The benefit of starting with a distribution is that intelligent configuration choices are provided as defaults for many of those parameters. Also, the order in which these components get deployed is very important, and the installation framework takes care of that.”

“And son,” my mother continues, “there are other things that distribution vendors provide based on their experience with customers, features like HA for the control plane and compute nodes, non-disruptive upgrade, and support for a wide range of hypervisors. And speaking of support, remember that Day 1 operations is a small part of the whole lifecycle of your cloud. What happens on Day 2 when something doesn’t work, or there’s a bug, or a security vulnerability that needs to be fixed, or goodness knows what else? If you’re going to put all those mission critical workloads including your family finances in your cloud, you have to think about how to keep that cloud up and running 24/7 and how you will continue to maintain it years after that.” My mother is wise like that.

“But what about vendor lock-in, Mom?” I reply. “Aren’t I painting myself into a corner by using a distro from a commercial vendor?”

“Ah, always such a worried boy. That’s a valid concern. To be called an OpenStack distro, vendors have to include the core projects – nova, neutron, cinder, swift, glance, keystone – but there are dozens of other projects that provide additional functionality. So a distribution vendor makes choices about which ones to include based on their readiness, supportability, what they see in the market, and what their customers are asking for. But when you deploy a commercial distro you don’t have to install those other projects if you don’t need them. So you can keep control over the makeup of your OpenStack cloud.”

“And finally,” my mother says, with just a tiny bit of chastisement, “you of all people should know that choosing an OpenStack distribution from an open source veteran like SUSE, who is building the solution on enterprise-grade, hardened Linux will also ensure that you get support for the entire solution stack. This includes the host and guest Linux OS and not just the OpenStack software layer. It also gives you the best interoperability, with support for just about any hardware that you are likely to have in your data center, or, um… your basement.”

Wow, my mother knows WAY more about OpenStack and Linux than I thought, and thanks to her persuasive argument, I am convinced that starting with an OpenStack distro is the right thing to do for most use cases. But I still have a few other OpenStack questions for her, and you’ll have to wait for Part II to see those.


Share with friends and colleagues on social media
Tags: ,
Category: OpenStack, Technical Solutions
This entry was posted Tuesday, 10 May, 2016 at 10:47 am
You can follow any responses to this entry via RSS.

Comments

  • nlsnow says:

    Thanks for this fun article!

  • Leave a Reply

    Your email address will not be published. Required fields are marked *