This year’s just-wrapped Hadoop Summit 2015–particularly the exhibit hall–was somewhat subdued compared to last year. Yet attendance was up at least 10% over 2014. Could attendees be suffering Hadoop-fatigue? Or just tradeshow-fatigue? More reasonable explanation: Hadoop’s not a novelty anymore, and as adoption has climbed, attendees seek out optimization counsel, when in the past they focused on education.
Simba partner Hortonworks’ president Herb Cunitz introduced the day-one keynotes, entering with exceptionally-well-choreographed, blacklight-lit, neon-illuminated dancers. He wore a blazer covered with glow sticks. (Interestingly, it’s what he wears every Tuesday.) A straightforward, yet enlightenting challenge from Cunitz: “Can you create a [viable] business model based on the insight you get from your data?”
For at least one vendor, Hadoop Summit 2015 was quieter than last year. (Hey, even flash mobs need the occasional day off.)
Fun fact: Hadoop Summit attendees arrived from 38 different countries. (Canada included, of course.)
In his day-one keynote, Hortonworks CEO Ron Bearden compared Hadoop adoption to RDBMS. (Guess which won the adoption-curve race.)
Simba’s own Sylvia Lee and Jenny Chu joined with some 40 other fellow Hadoopers on the third annual Hadoop Summit bike ride. Sure, they had to get up early, but they got a nice shout-out on the day 2 keynote. (Look for the Simba logos in this keynote-screenscraped pic.)
Best (though admittedly cynical) buzzword of the week (heard from an analyst): “Feature plaque” — little-utilized legacy features that hang around in older RDBMS products because the RDBMS developer has to keep that one long-term customer happy. AKA “abandonware”.
Same analyst noted rather counter-intuitively that in some cases, Hadoop innovation can actually driving a wedge (and foster conflict) between DBAs and developers. Data formats and the data itself are growing exponentially. Successful Hadoop deployments–regardless of scale–make peace between analysts and IT leads by delivering scalability, offering extensibility, and perhaps most importantly, reducing complexity across disparate platforms.
Cool technology: Arcadia Data. Arcadia’s technology puts the visualization where the data is. (And no, their marketing tagline doesn’t have a dangling participle. That’s my grammar, not theirs.) I envision it as “verticalizing” data preparation at the site of the actual data. Next objective: Building in scalable extensibility via ODBC and JDBC connectivity.
The Los Angeles Lakers attended Hadoop Summit. Really. Well, not Kobe, but their data guy made it. The Lakers are known for their Sabermetrics-like data analysis of player performance. Golden State Warrior fans in attendance were quick to note that Lakers’ data models might need some tweaking after last season’s flameout.
Disney has some exciting Big Data initiatives in play–including Spark and Hive. Disney software engineer Caleb Jones’ speech called out Disney’s reliance on ODBC and JDBC connectivity for Hive and Spark (that just happens to have been developed by a big little company in Vancouver, BC, Canada).
For MapR, Hadoop Summit 2015 was all about Drill…specifically, the MapR Hadoop distro that now features Apache Drill (with, I should add, embedded Simba ODBC connectivity). One presentation highlight was Tomer Shiran (MapR VP of Product Management) demoing Drill with Jacques Nadeau of the Apache Software Foundation.
Finally, what was most evident at Hadoop Summit 2015 was how far Simba Technologies has come in the Big Data market sphere. Nearly every single exhibitor uses some flavor (in some cases, many flavors) of Simba connectivity for Hadoop data access. In addition, the on-the-floor conversation has evolved: The event featured an educated, Hadoop-expert audience, and our engagement with visitors to the Simba booth was at a deep technical level. Next year? Maybe we’ll take the next step to on-site driver prototyping!