Published at 23.10.2016
Table of Contents
Here we are again – after the Cloud Foundry Summit USA in Santa Clara in May, the community met again in Frankfurt for the Cloud Foundry Summit EU in September. As a silver sponsor the anynines team did not only present a booth but was also involved in organizing the Unconference. In this article, we will share some impressions of the conference.
As already mentioned, the anynines team was involved in organizing the Unconference. To be accurate, it was Sara Lenz who organized it together with Paula Kennedy from Pivotal and they did a very good job. Anyway, this and the fact that some of the anynines engineers gave training sessions, was the reason why we already arrived on Monday in Frankfurt, one day before the actual conference started.
After the introduction speech, some of the sponsors had the opportunity to give Lightning Talks. Among others also our CEO Julian Fischer. The Lightning Talks were followed by some discussion rounds and a little dinner with time for networking with the other sponsors. Altogether this was a nice start for the actual conference and we were excited for the next days.
On the first day of the actual conference we were busy setting up our booth and preparing all the giveaways. Luckily we were offered some fruits and breakfast to energize ourselves :).
While preparing everything, keynotes already started in the main room, but I couldn’t attend, as I was busy preparing everything. After we set up everything we had a look at the other booths and connected a little bit the other sponsors.
For me, it was the first CF Summit and unlike some of the other anynines team members, I almost didn’t know any of the other companies. That’s why I checked out what the other companies were doing and what giveaways they had. The most impressive giveaways were given by the grape up team, as they had grapes to give away, that was awesome.
After some talks, the first break took place and the attendees then began to visit the booths and so some of us started to talk with them about our offerings, while the other ones visited some talks. For me, there were no interesting talks for the first part of the day, so I first helped out at the booth.
Then the second part of the day, I wanted to visit some talks and the first one I attended was held by Duncan Winn, who works for Pivotal. He talked about simple abstractions for the challenge complexity regarding the Cloud Foundry architecture.
Therefore, he first talked about why Pivotal has refactored the DEA architecture into the new Diego architecture. After that, he talked about almost every component or group of components of Cloud Foundry and explained how the complexity is handled by building strong abstractions between them. During the talk I had the feeling that he must have been involved in the whole refactoring, because he knew a lot about the architecture.
The second talk I attended was some kind of an experience report about the Cloud Foundry Dojo held by Cornelius Schumacher from SUSE Linux and Michael Trestman from Pivotal. Pivotal treats dojo attendees like new applicants, so you first have to do a remote pairing session with one of the Pivotal engineers in order to see if the technical and social skills fit to Pivotal.
If you passed this session, you are part of the dojo. A working day at Pivotal starts with a breakfast followed by a global standup. After that, they do Pair Programming all day every day and they rotate every 1-2 days.
They say Pair Programming in general leads to better decision making, higher quality and a propagation of Knowledge. That is why Pivotal sees no need for Code Reviews, which I think is quite interesting.
“There are business processes and development methodologies unlike software but they’re running on the wetware that’s in our heads.” —Daniel Jones, EngineerBetter
The last talk for this day was titled “Anthropic Sympathy – How Cloud Foundry & Continuous Delivery Make Humans Happier and More Productive” by Daniel Jones from Engineering Better. Daniel basically talked about why an agile DevOps culture doing Continuous Delivery and delivering value in small iterations is better than a conventional waterfall process with the typical silos (like separated Dev, Ops, Database teams).
He showed, how Cloud Foundry supports that and he underlined his statements with psychological facts. For me, this was the most interesting and impressive talk of the whole conference.
On the second day there was not as much traffic at the booths as on the first day, so we had little bit more free time. So the first talk I attended on that day was about Docker & Cloud Foundry.
In his talk Alexandre Vasseur, working for Pivotal, talked about 3 Use Cases for using Docker with Cloud Foundry. The first Use Case is pushing an app to Cloud Foundry as a Container.
As a developer you have to build your container and push it to the Docker Registry and then you can push it to Cloud Foundry. The second Use Case is to create backing services via Docker containers, whereat Alexandre mentioned that this is more of a R&D project.
Anyway, this Use Case describes how backing services can be created where the service broker utilizes Docker and Docker Swarm in order to provision a backing service. The last Use Case mentioned by Alexandre was not directly connected to Cloud Foundry, as it described Docker containers for pipeline tasks.
He talked about Concourse.ci and described how basically every step in Concourse is performed by building Docker containers and performing tasks on them.
— Ben Wilcock (@benbravo73) 28. September 2016
The second talk for that day was titled “Undo for Service Brokers” held by Justin Carter & Dr. Nic from Stark & Wayne.
They introduced a Cloud Foundry Plugin to recover deleted service instances. Basically the store relations between uuids and names of each service because the Service Broker only works with uuids.
Second they store backups (Database dumps, etc.) in a S3 Bucket.
Now, if you want to undo a service, the plugin resolves the uuid for a given service name, creates it again and restores the dump. The Plugin seems to work with MySQL and PostgreSQL but not with MongoDB as far as I got it. Anyway, I think this is a pretty cool plugin.
The last talk for me on that day was also held by Justin Carter but this time together with Xiujiao Gao also from Stark & Wayne. They talked about Subway – a tool to scale Single-Node Service Brokers. Basically, Subway is kind of a proxy that allows you to scale out Service Brokers. Therefore, Subway checks which Service Broker actually is able to handle the request in order to create a service. It is deployable as an App or as a BOSH release.
This is also really nice tool, if your service broker cannot scale by default.
Altogether, the CF Summit Europe, which was my first CF Summit, was very interesting and a really good experience. A lot of impressions, a lot of interesting people, companies and conversations.
I don’t know exactly why, but seeing all the people of the CF community and listening to some of the talks was very motivating.
It was interesting to see how big the interest in Cloud Foundry really is, also for big companies that now are starting to move in direction of a DevOps culture and Continuous Everything with the help of Cloud Foundry.