Node Interactive 2015, the first conference run by the nascent Node Foundation, wrapped up yesterday in Portland. The name of the conference was apt, because many people remarked on just how much real interaction there was, in contrast to some events where people keep their heads down and stick with their own groups. Here, there was lots of lively discussion and interaction after sessions, during breaks, on the showcase floor, and at the evening parties.
The level of excitement, engagement, and anticipation reminded me of the first JavaOne conference that I attended almost 20 years ago (yes, I’m an old-timer!) With over 750 attendees, Node Interactive 2015 was one of the largest Node.js conferences so far, and looks to become the “must attend” conference in the Node world, as JavaOne is in the Java world.
In this blog post, I’ll briefly summarize some of the most popular and talked about sessions.
One of the most popular talks on the first day was Debugging Node.js in Production by Yunong Xiao, a Node engineer at NetFlix. Among the issues he addressed was how to sample stack traces from a process without affecting the performance of the process, and how to use flame graphs to find the one line of code (out of millions of lines) that is causing performance issues. He said that the results of debugging using such techniques were dramatically reduced request latency, reduced CPU use, and increased throughput. I would be remiss if I didn’t mention that StrongLoop Arc provides flame graphs via its Tracing module, currently in early release.
Another popular topic in general was package and module management. Several sessions addressed this perennial pain point, including “Stellar Module Management” by David Dias of Protocol Labs and “Getting a Handle On Your Dependencies” by Dan Silivestru of bitHound (watch a video of his presentation).
A popular talk on the second day was “Node.js API Pitfalls: Can You Spot Them?” by Sam Roberts of Strongloop/IBM. Sam engaged developers with a series of questions on small code snippets, with surprising (and sometimes perverse) answers. See the blog post from my IBM colleague Sai Vennam for a summary, or watch a video of the presentation. To view the slides for this presentation:
git clone https://gist.github.com/sam-github/4c5c019b92cf95fb6571
Then open the index.html file in your browser.
Another session on the second day that provided great insights was the panel discussion “Node.js for Enterprise APIs,” with representatives from Bank of America, TD Bank, and GoDaddy, moderated by Juan Carlos Soto of StrongLoop/IBM. The panelists spoke about their experiences “in the trenches” using Node in enterprise environments. All three said they are actively deploying APIs written in Node.js.
Both of the panelists from banks noted that their industry requires indemnification for liability reasons, which can make open-source technology problematic. For this reason, they look to companies such as IBM/StrongLoop that can stand behind and support OS software. They also said their development teams are starting to move away from the old “waterfall” development process to Agile development, and Node is a great fit for that because it’s well-suited for rapid prototyping and development.