To kick off my first month at Pivotal, I took a tour of some European tech conferences with Andrew Shafer. Why does that “kick off” my first month here, good question. We went to the RedMonk conference, Monkigras, in London; then FOSDEM in Brussels; and finally Config Management Camp in Ghent. Some of the key themes driving the use of more software in businesses (the whole “software is eating the world” jag, if you prefer) did pop up frequently, however, which was nice to see out in the European IT soup. What was most exciting to see was the passion speakers and attendees at each conference had around software.
Monkigras: Nordic Culture
Billed as a celebration of “Nordic Craft Culture and Tech” this year, the two-day RedMonk conference exactly that: tales from the Nordic countries sprinkled with IT-land parallelisms. Having worked at RedMonk myself for some years, I pretty much knew what to expect: hearing about novel uses of new technology in both the real world and our computer-land bubble.
One talk stood out in my memory in particular as a great discussion of software trying desperately to inject itself into everyday life.
Helena Bengtsson, Editor of Data Projects at The Guardian newspaper, told us about her adventures in data-driven journalism. That is, the seemingly simple task of using a tiny bit of coding and data analysis to mine the world for stories. In particular, Helena was interested in looking into various government activities, for example, seeing how often each lobbyist visited government officials.
Each of the data-driven investigations quickly turned into a sort of cat-and-mouse game where the Helena would find a piece of interesting data and then the government would slightly change how that data was made available in order to, one presumes, add just enough chaff to the system to make Helena’s task more difficult. For example, Helena asked for a dump of all lobbyist visits and received a very tidy CSV file for the past 30 days. As she asked for more complete updates, the government changed its reporting mechanism to just be daily snap-shots, expiring any data older than 24 hours. So, being tenacious, Helena simply asked for these snap-shots daily. There were other nightmarish ETL tales like having to handle PDFs and other ill-formatted data.
What I liked about this talk was how it showed how us California Ideology types quickly run into the rotating knives of real life. There’s wondrous things to do with data analysis—even simple CSV files in Excel, Perl scripts, and Access databases, let alone the destriers of proper Big Data rigs like Apache Hadoop®—but the meat-ware layer of people and their policies are the real gate-keeper to spinning gold out of data dross.
Speaking of spinning, Ankle Holst discussed knitting, which really had very little to do with tech, but afforded the chance to show off some most excellent Icelandic sweaters. I need to get some of those.
Compared to Monkigras, or any normal tech conference for that matter, FOSDEM is a sprawling octopus lost in a well-chilled sea of concrete and beer. Which is to say, it’s a jam packed conference with the nerd dial on maximum. The conference started 14 years ago and is run in a very “free and open source” style with no formal registration and volunteer driven tracks. It can all seem chaotic and ill-run (for example, the talks all overlap each other in timing and rooms are often full to capacity), but there’s an excellent variety of topics to seek out and discover: many arms to that octopus and all.
Down at the BOSH-layer of the stack, we spent much time in the configuration management track hearing about a wide range of configuration management tools from the stalwarts of Puppet and Chef, to Juju, to newer takes like Consul.
The Internet of Things track was also interesting in that it zoomed way down the stack to the device code level. We listened to several sessions that followed the standard IoT tech talk: we have this light bulb, a sensor, and a little device that emits sounds…and look how my code makes the light bulb change colors according to how much BO the guy sitting next to you is emitting! OK, maybe that was how I interpreted the demo playing out based on where I was sitting…
More seriously, what you saw playing out in the IoT track sessions was development of the very lowest levels of the IoT stack: simply coordinating the flow of information between all the devices in a IoT network. To me this points towards all the complexity we’ll have in a IoT-driven world. While we may think “internet scale” is difficult and the providence of “unicorns,” you can start to feel that IoT technology is even more complex and the domain of winged unicorns, perhaps with 7 horns each and monstrous appetites for hard problems.
A recent Gartner survey pegged this market as pretty early, but still had 60% of respondents saying that IoT would have positive effects on their business in five years, a pretty quick clip from blinking light-bulbs to mainstream business utility. There’s clearly a lot of interesting work to be done between now and then.
I’ve been interested in how Pivotal can help in the IoT space and I’ve had the chance to see how some of our customers are looking to use Pivotal Cloud Foundry as a platform on-top of all the complexity of IoT. Part of what’s appealing is getting the application development platform standardizing benefits that come with using Cloud Foundry. This is key, to me, because just as with the classic Internet, it’s clear that the benefits of IoT will be creating an open ecosystem for developers to come and dream up new types of applications across all the connected devices in IoT scenarios. Rather than shaving the yak of creating a new application platform each time, I’m hopeful that Cloud Foundry can shave the yak once and let developers focus on the actual applications, from dimming light-bulbs to better tuning jet engines.
Configuration Management Camp
After loading up on frites in Brussels we headed to Ghent for a smaller conference, Configuration Management Camp. The difference between this conference and FOSDEM was striking: #cfgmgmtcamp was much smaller and in the facilities of a tidy college, no chilly cephalopods here.
In recent years, there’s been a huge upsurge in enterprise interest in this space, and rightly so. The early movers who glommed onto Puppet and Chef achieved all sorts of Agile advantages when it came to taming their cloud and good old fashioned IT usage. And they still do, obviously.
There were many tracks and sessions going over the stalwarts of configuration management from Puppet, to Chef, Salt, Ansible, and the work coming out of newer places like HashiCorp and ourselves here at Pivotal. The theme I liked the most was a sort of “beyond configuration management” idea. The two opening keynotes on day one both hit on this theme, one from Mitchell Hashimoto (which was itself titled “beyond configuration management”) and the next from Andrew.
It’d be easy to think such a theme was along the lines of “configuration management is dead!” but it was far from that in my books. Instead, I looked at these talks as urging the crowd to start concerning themselves with more than just configuration management and look further up the stack to the application enablement layer and think about how configuration management tools can keep making life better further up the stack.
Also, if you like the whole listening thing, I snuck in an audio recording of his talk if you’d like to take a gander.
This is all being recorded
My summaries are a pretty dim shadow of the actual events. Thankfully, if you’re interested most everything at these three conferences was being recorded. FOSDEM has archives up, and you can see last year’s Monkigras videos as well. Keep your eyes peeled for the 2015 videos, there’s lots of good stuff in there.
Editor’s Note: Apache, Apache Hadoop, Hadoop, and the yellow elephant logo are either registered trademarks or trademarks of the Apache Software Foundation in the United States and/or other countries.
About the Author