OpenStack Summit Berlin 2018: making the right connections

With OpenStack Summit Berlin 2018 now behind us, I thought I would share my views and experiences of the event. It’s always a highlight in the calendar, and this year was no exception, providing great insight into current trends, showcasing the amazing OpenStack community, and, of course, giving everyone the chance to pit their wits against Rackspace’s Fix My Stack challenge!

As expected, City Cube in Berlin was a great venue for an event of this scale, and everything kicked off with the usual round of keynotes from sponsors. A key message that came through was how the ecosystem is evolving, and that we are seeing new workloads being deployed that bring with them new hardware requirements. To that end, this was the last OpenStack Summit that will be held, with Denver 2019 marking the start of Open Infrastructure Summits to encompass a multitude of open source technology. OpenStack is at the heart of this.

Machine learning, AI and edge computing were all cited as big growth areas, driving an increased demand for GPUs and enhanced network control. Platforms powered by OpenStack need to be able to bring all these together on one easy to operate platform. Just installing Docker is no longer enough; instead, we need effective solutions that can provide users with the data and features they need in today’s fast-evolving cloud world.

Using the right cloud 

We heard from VW how they deploy their whole infra as code. To achieve this, they switched from Open vSwitch to Open Contrail, the open source software-defined networking stack from Juniper. They said networking problems “are the hardest to troubleshoot”, though, and so it would seem there is still some work to be done to simplify the utopian dream of a ‘data centre as code’!

VW also reiterated a fact I think a lot of people keep missing: that operating a private cloud, when done correctly, is significantly cheaper than using the large public cloud providers – not to mention the benefits you get in performance, control and security. The OpenStack community as a whole is a HyperScaler, and is competing with the public cloud HyperScalers.

Throughout the conference, cloud operations was another common topic, from tooling for the initial deployment to how to best handle upgrades and migrations. During a panel discussion hosted by our own Justin Shepherd, we heard from Pipedrive CTO Sergei Anikin about the importance of mobility and being able to easily upgrade a production platform with no downtime – something Pipedrive achieved thanks to their Rackspace Private Cloud deployments.

EdgeVerve also shared their experiences and explained how the performance of their private cloud (hosted on commodity hardware in their DCs, but also operated by Rackspace) exceeded that of competing virtualization and public cloud environment tests that they had performed. Suman Mukhopadhyay, EdgeVerve Senior Director & Head of Technology Infrastructure/Cloud, had expected a reduction in performance when transitioned from public cloud, but in fact achieved significant performance gains in the compute and storage layers. These efficiencies are now driving further adoption and growth at EdgeVerve.

The future of OpenStack

On day two, we learnt the next release of OpenStack will be called ‘Train’ – a nod to the developers being kept awake by trains at the last Project Team Gathering (PTG). The first Open Infrastructure Summit will be held in Denver, 29 April–1 May 2019 and will be co-located with the PTG again.

The growing move to Open Source was also regularly highlighted, with an ever-increasing number of companies contributing to open source projects, and not just using the code.  For example, in the past 12 months there have been over 70,000 commits to OpenStack, placing it in the top three most active open source projects on the planet. Among those changes was the move to Nspawn for OpenStack-Ansible. For those not familiar with this project, it was originally developed by Rackspace and is now used to deploy OpenStack Private Clouds around the globe.

At the Rackspace booth

Kubernetes and GPU were also a common topic of conversation, with my Rackspace colleagues having numerous discussions on our stand – a buzzing hive of activity for the duration of the summit. This was helped by our awesome authors who were giving away hundreds of signed books, resulting in long queues. We had Kevin Jackson, Cody Bunch, Egel Sigler and James Denton signing the latest edition of the their OpenStack Cookbook, and James Denton signing his OpenStack Networking ‘bible’.

We also ran a new technically advanced version of our Fix My Stack Challenge, which proved as popular as ever. After a hard-fought battle, with three finalists having to first work their way through two rounds based on an OpenStack-Ansible AIO deployment, they then had to up their game and fix a solution based on my very own OpenStack-Multa deployment. Congratulations to Elvis Noudjeu (centre) who won a HP Chromebook for coming first, and to Darryl Weaver (right) and Cristian  (left) who came in a very close second and third. Each of them took home a pair of Omen Headphones.

In closing…

Everything about the event – venue; keynotes; presentations; official parties – was a success. My colleagues and I had many interesting conversations with existing OpenStack users and organisations looking to take their first steps into the exciting world of open source cloud technologies.

Containers, AI, edge computing and machine learning appear to be the big growth areas; but although these new exciting techs grab more column inches, we need to be mindful that they still rely on underlying infrastructure – the very infrastructure you need to deploy, upgrade, expand and enhance on an ongoing basis if you are to be successful on your open source cloud journey.

If you’re interested in what OpenStack could do for your company, or want to get more from your existing OpenStack deployments, please get in touch:


Please enter your comment!
Please enter your name here