In the past week, I have had so many opportunities to discuss private cloud with our customers, and I am becoming very optimistic about the opportunities being presented to me. I came home with a ton of enthusiasm for OpenStack clouds and I'm just keeping that momentum going. That said, I'm going to provide a detailed review of my OpenStack Summit experience here over multiple posts. This is, in part, to share my love and joy for the OpenStack community and what I experienced in Portland - and also for my own selfish joy of getting to re-live the experience all over again.
I arrived in Portland late on Sunday night and flew in on a plane that was full of stackers. After a horrible night of sleep thanks to being in a hotel right along the 1st. street train line, I decided to walk into the city and get myself a green smoothie for breakfast before walking across the Steel Bridge and just over a mile to the convention center. It was a nice morning out, and I like the idea of walking around a city I haven't been in to get a feel for it. I used Yelp to find Kure, ordered myself a delicious green monster for breakfast and walked on over to the convention center.
The line for registration seemed to go on forever - until I figured out I was in the wrong line. I found the right line and got checked in by a team of guys with iPads running some custom app. I wonder if it was powered by Nova Compute? After some wait, my badge printed, I was handed my OpenStack hoodie (which has become like my security blanket) and off I went to my first day of sessions.
First up was "Getting from Grizzly to Havana: A Dev Ops Upgrade Pattern" which was very well presented by Rob Hirschfeld (@zehicle) from Dell. But first, an interlude - Rob started off by having the group record the chorus to his upcoming hit parody of Pawn Shop based on OpenStack (http://t.co/rX8V8WqMAc) .... Then he went on to talk about Dell's Crowbar management toolkit which can be used to manage an OpenStack deployment on Dell reference hardware. Since I had been asked this question earlier by a customer, my favorite takeaway from that session was "Upgrading from Essex to Grizzly is probably not a good idea." The real focus of the session was on using the layered DevOps approach to managing infrastructure as being a key component to upgrading. In a layered methodology, configurations are applied first to bare hardware, then at an operating system level and, at each step, additional custom configurations can be applied. This allows for a more granular approach to managing infrastructure and all of its components using a organized, systematical approach which is much simpler to manage than using legacy methodology. More focus is spent on delivery of platforms and applications, which is where real business value is realized.
Today, we don't have a clear upgrade path as OpenStack moves from one release to the next, but the need for this type of support has a tremendous effect on OpenStack's acceptance in the enterprise. I've been working on a variety of these solutions, and I just find that there are features in one platform that others are lacking - and vice versa. What if you have a truly commodity infrastructure? Which tool set do you invest in? This will all be discussed later in a session by Jason Cannavale from Rackspace...
No comments:
Post a Comment