Spring has sprung though I can barely believe it. Though as I walked along the Brighton seafront, with a bracing chill wind in my face and a red sunset at my heels, it did feel a little more spring-like.
I was in Brighton to attend my second, though really my first FLOSSUK.(1) I have had the fortune to be asked to speak on the current state of the Perl world in regards to developers, culture, environment and practices. At the same time I was to have a peek at some of the more modern tools.
FLOSS Spring is an event that, in popular mythology, is mostly attended by systems administrators. I feel this is a historical precedent, or maybe just a popular perception, that doesn't match the character of either the attendees or the phenomena. The talks, people and passion maybe all about modern techniques in DevOps, and a significant percentage are systems people. However they are also interested in languages, projects, development as well as systems. One of the best features is born from the fact that this conference is language agnostic. If it is open, it is wanted. And most likely respected, accepted, loved and discussed.(2)
Though a large number of the attendees may work in system engineering, they are also closely connected to development and deployment. This is perhaps the reason for the eponymous DevOps association. What we actually have is a broad sector of senior system managers, network architectural engineers, developers, system administrators and project managers. This is a great environment for cross-pollinating ideas, techniques, software solutions and problem solving.
The Shadowcat Team arrived a day early on the Tuesday to provide a last minute tutorial after one of the trainers dropped out due to unforseen circumstances. Ian Norton and Tom Bloor gave an introduction to Perl event, a variant on the presentation Ian has given on an number of occasions. This was attended by five students, one of whom specifically swapped to be on the Perl event and seemed most keen. There was a general buzz in the room and it was good to see our latest 'minion' hold his own teaching others.
The conference started in earnest on the Wednesday. The keynote speech on the changes to the UK Government website was enlightening, especially if you deal with site design and delivery. The approaches, changes and entire revolution of ideas was fascinating to learn.
I also really liked the talk by Jans Mens on Ansible, unfortunately it quickly lost me and I think I managed to confuse his section on Roles with the ideas of Roles from Computer Languages, with particular reference to how they are used in Moose and Perl6.
I enjoyed Simon Riggs' talk on PostgresSQL. Postgres seems to be a fast-moving database system. I know a lot of effort has been expended on noSQL of recent years but Postgres has some real power and features (being able to run eighty-five million request an hour on a low powered laptop being just the thin edge of a large wedge) and a yearly development cycle that keeps a high momentum.
The Thursday started with an almost UK Open Source conference tradition. Bytemark, once again, supplied a personalised conference gift in the shape of a mug. This had a unique message, conference branding and your own name.
The first talk I heard was given by Bernd Erk on OpenNebula. OpenNebula is a datacentre virtualisation system, however think more that it creates ‘private clouds not public clouds’. OpenNebula however is all about small clouds for internal company usage. A highlight is that it works with a range of open tools to manage configuration, replication and reporting.
Chris Jones, gave an interesting talk about OpenStack. The latest move in OpenStack is called TripleO. The short, and mostly true version of this, is to be running a cloud with VMs and attached storage using OpenStack on OpenStack. Hence why they are calling this Triple-O, OpenStack On OpenStack you see, almost alliterative. The idea is to use thesame tool to deploy as the tool that is being deployed, as many of the components to do this are already available inside that tool (and now I get lost in recursion).
Conceptually they see it as a cloud on top of a cloud. This is not strictly true, but it helps as a conceptual model to understand what is actually happening. It gives them the name of undercloud - the abstracted management layer that sits on the hardware; and the overcloud which is the client facing open stack implementation.
The second talk of the day given by Bernd Erk was a quite interesting discussion on the roadmap for Icinga2. Icinga is a fork of the ever-popular Nagios and is compatible with the modules from the Nagios eco-system. It was good to hear Bernd talk with a lot of enthusiasm for the current state of network monitoring and cloud based PaaS.
Matt S. Trout, of Shadowcat infamy, gave a talk on Prolog and Devops Logique - as always Matt gave a wander through the recesses of his mind, the history of computing and system approaches, with various conversations concerning configuration of servers.
On his journey Matt assessed the manner in which a number of other languages have implemented or addressed some of the issues he is facing. This approach allows him to evaluate and understand the argument in a much broader manner. The end goal seems to be to handle dependencies and to manage them well on a system perhaps utilising the Perl scripting language.
David Griffith gave one of the best talks of the day on Test Frameworks and the issues they have encountered and overcome at Durham University. It was strange to listen to a talk that was both complimentary on the achievements yet also modest and reflective. David was speaking about the cultural changes in remote workers and cloud systems, which nicely closed the event for me as we started with a keynote that emphasised a revolution in the culture of a department.
The event itself was closed by Kimball Johnson who presented Josette from O'Reilly with flowers as this is her last conference as an O'Reilly representative. This was a deep sadness for most of us at the conference. Josette has been a great treasure to this event and many other open source communities and she will be greatly missed as O’Reilly’s voice in the UK and European community.
Kimball then presented awards for the best talk, best lightning talk and several honourable mentions. I was greatly honoured, and proud, to receive the award for best lightning talk and an honourable mention for my longer talk. The best prize however was simply being at the event and enjoying the talks and the people.
A few other thoughts have to be conveyed before I close this piece.
One of the major highlights of the week was the 'Spaceship' that was built, created and manned by the chaps at the London HackSpace. They brought this wonderful experience to the conference dinner and teams of happy victims were led into the event over the course of the night. I think I would quickly run out of synonyms if I tried to write a range of superlatives on the brilliance of this experience. The spaceship looks amazing, the various effects, screens, buttons and stories have a lot of thought and care gone into them. The LHS crew both run the simulation and act out the varying storylines with the crew and even dress for the occasion.(3)
I would like to thank the conference organisers, and especially Kimball Johnson, for inviting me to talk at the Spring Conference and for asking me to write on my experience which you can read in the UKUUG newsletter and on their website.
(1) Though the term might be thoroughly soaked if they live in the UK.
(2) For that fact alone we should be shouting to all the other language groups and saying come one and all, attend, mix and learn, be a part of an open forum not a closed view.
(3) For a better description of crew see willing participants, attendees or as they are also known, victims.