The recent OpenShift Commons event at London’s QEII Centre was a huge success. Over 200 attendees enjoyed talks and panels on the state of container technologies, the future of OpenShift and forthcoming features, the OpenShift Service Catalog, and the use of OpenShift in government and the financial sector. Secnix were proud to be sponsors and co-organisers of the event.
RedHat’s directory of community development Diane Mueller was the principal RedHat organiser and she opened the event with class! You can watch Diane’s introduction here.
As Diane explained it, the growth in the OpenShift community, with 40+ new members in the last 100 days alone, demonstrates the rude health of collaboration both upstream and across streams. With over 330 organisations in over 50 countries, and code contributions from at least 70 of those organisations, plus a new community model and improved communications between RedHat and collaborators, OpenShift Commons goes from strength to strength.
It was also Diane’s job to make sure people stuck to the schedule, though she had a tough time on her hands with the irrepressible Dan Walsh, RedHat’s consulting engineer. Dan’s keynote speech was on the state of container technologies and if he ran over a little it was only due to popular demand. You can watch Dan’s presentation below.
RedHat’s goal is to make containers as generic a term in infrastructure as PDF is for documents or Linux for the backbone of half the world’s software. Dan delved deep into container technology, introducing Skopeo for the viewing of remote registry manifests, fetch image layers, and the importance of runc to the future of container technology.
Dan also explained that what Kubernetes needs to run a container is a new, lightweight daemon. Handily RedHat are ready to unveil just that in the form of CRI-O – an OCI-based implementation of Kubernetes Container Runtime Interface. This should mean reduced need for Mesosphere, Swarm, Docker and the rest as Kubernetes combines seamlessly with CRI-O, which will share a version number to simplify things further.
You needed your wits about you to keep up with the next speaker. RedHat product manager Diogenes Rettori zoomed through a battery of new features in his talk on the future of OpenShift 3.x and at times it was hard to keep up. Could have sworn he mentioned dog food at one point.
Dense cluster scaling, bringing Prometheus cluster to OpenShift, new features in metrics and autoscaling, and a new beta metrics API for pods and nodes, replacing heapster. Diogenes explained how ‘extensibility everywhere’ is the key to OpenShift’s future with improvements to service load balancing, beta IPVS kube-proxy implementation and a whole load more. You can watch Diogenes’s presentation here.
A set of high-quality lightning talks from the event’s partners, such as nearForm, CoScale, and StorageOS took us to the lunch break. After a splendid lunch – of fish and chips, no less – principal software engineer at Red Hat Paul Morie took over for a hands-on demo of the OpenShift Service Catalog.
The goal of the Service Catalog is to remove manual, error-prone processes that can occur when a service consumer looks to acquire a service from a provider. Paul explained the updates and future features coming to the Service Catalog to an audience eager to exploit its full potential, very much including Secnix ourselves. You can watch Paul’s presentation below.
In a system utilising the Service Catalog, an engineer or operator may need access to a complicated set of infrastructure such as a database with direct attached storage, perhaps for load testing. Instead of submitting a ticket and waiting, a query to the Service Catalog lists managed services provided by service brokers, defined by the Open Service Broker API spec.
The operator creates a ServiceInstance and the service broker provisions the service. After provisioning, the operator binds to the service by creating a ServiceBinding. The broker does the work of creating the binding, possibly by creating a new set of credentials and returns the connection information.
The operator updates the DeploymentConfig to consume the secret created by the Service Catalog. This invokes change triggers, which react by updating the ReplicationController in turn automatically terminating existing pods and creates new pods with the injected information.
Finally, Paul showed off a new project called the OSB starter pack that is aimed at giving developers a quickstart for creating service brokers that implement the Open Service Broker API. If you need to implement a service broker at breakneck speed, this is the place to be!
That was the plain English explanation – rest assured Paul explained it a lot better than us!
The day was rounded off with two panel discussions. The first was on OpenShift use in the financial services sector; perhaps unsurprisingly the camera was turned off for this part as Barclays, CYBG, and Citi keep their OpenShift cards close to their chests in a fiercely competitive industry.
The second panel involved OpenShift’s use within the UK government. Secnix have overseen successful projects at the Department for International Trade and Department for Education, but adoption remains sporadic. The panel discussed how to win over skeptical officials to the OpenShift cause given the great potential for new tech to increase efficiency and reduce costs in areas that consistently need more for less. You can watch the panel discussion here.
Serverless computing came up multiple times during the day and it will be interesting to see how things progress in that area over the coming months. Attendees were also digesting the news that RedHat has agreed to acquire CoreOS, but that’s a story for another day.
Our own Justin Cook was the driving force behind the OpenShift Commons event, having set its wheels in motion at KubeCon in December. In just a single month, Secnix, RedHat and other partners managed to pull together this hugely popular gathering in the heart of Westminster, and Justin was roundly applauded for a job well done (not least the ample beers at the end).
It just goes to show the kind of will to win you’ll be enlisting when you hire Secnix for your project.