The week is coming to completion here in Paris. The OpenStack Summit season for 2014 is completed, and the last few hours of the Kilo Design Summit are finishing up tomorrow. I was even lucky enough to be able to meet today with OpenStack Foundation COO, Mark Collier, and this was the perfect way to round out the week for me. I had a really fun week with the vBrownBag team here and all of the great people who were at the conference to celebrate this next phase in the OpenStack timeline.
Less Cheerleading, More Working Together
The theme I picked up from a lot of the conversations around the event, and from some of the sessions I watched, is around what we need to do as a community to increase the viability of the OpenStack ecosystem. It seems odd to say “less cheerleading”, but what many have noticed from some sessions in the past is that there was always this strong message of “OpenStack is great, so you need to stop doing IT the way you are doing it today, because this is better”.
I love what OpenStack has done up to now, and I love where the whole project is going. It feels like we are crossing the point where hype is being outweighed by history as more customers tell their success stories. We have a massively growing customer ecosystem across all verticals, and in every size and shape you can imagine.
Vendors, developers, and all of the contributors are bringing the challenges with OpenStack to the fore, proudly. The acknowledgement of some of the difficulties in the projects has become the new conversation. We are working together more tightly, and more contributors are diving in to attend to requirements and challenges with features and fixes. There are a few more challenges ahead still at this point.
Upgrades Need to Be Easier
Unfortunately, we can’t all stay on trunk and push to production with every update that happens. This means that the process for upgrading components between releases has to become more fluid and simple. There are great strides being made in various programs to ease the pain for upgrades. Discussions around LTS (Long Term Support) editions is also underway which could help on this front potentially.
Swift has the advantage of being a distributed system that allows for upgrades to be done while the system continues to operate. We hope to see this sort of capability built into the other programs in some way. In-place upgrades or migration upgrades will be a welcome change for many customers who are having concerns about the long term operational processes.
One word: Neutron
The Neutron program has been going through some growing pains since it was made an integrated program in Folsom. Just like when Mark Twain wrote “the report of my death was an exaggeration”, we have seen the life of nova-network extended and celebrated. As Neutron evolved through the Grizzly and Havana releases, we saw the positive changes including the introduction of the ML2 (Modular Layer 2) plugin framework implemented.
Neutron has been challenging operators with scale, upgrades, and also configuration. Development and strengthening of the Neutron program is happening with many great contributors that marks some positive moves happening. There is still a lot of work ahead.
Many Enterprises Love High Availability
It’s not about pets and cattle. It’s about legacy applications that are in need of redesign, but a long cycle of development that will be outpaced by hardware and operating system refresh. Clouds aren’t scary, legacy processes are. In the interim we have to accommodate the workloads according to the real business requirements.
Have you ever been told that a storage product could get you 25:1 compression, but when you put your data on it the compression only reaches 12:1 and the vendor says “Your data is the problem”? We can’t let this happen to the OpenStack discussion. The customer workload is where the business requirements are derived from. There is no such thing as a bad workload. These are simply new user stories and potentially edge cases that have to be dealt with.
It was a great week all around and I hope to bring much more OpenStack Summit content to the community over the next while.
You can watch the entire set of sessions that were at this year’s Paris summit on YouTube at the OpenStack Foundation channel. This is the full playlist, so feel free to bookmark here and watch all the OpenStack goodness!
1 thought on “OpenStack Je T’aime: Another Great OpenStack Summit”