Observations from Gartner Data & Analytics Summit – Grapevine, Texas, March 2017

Contributed by

6 min read

Gartner put on a most delightful Data & Analytics Summit in Grapevine, Texas, the week of March 6, 2017. One could easily call this the Gartner Big Data Summit, but I think Gartner’s approach here – to combine data, analytics, new applications, master data management, open source, and more – reveals the breadth of what organizations are thinking about, planning, and doing with their data. While it’s fresh, I wanted to comment a bit on three tweets that I sent out last week about the show.

Most importantly (in my eyes), these tweets are about conversations I had with customers, prospects, and practitioners – not with analysts or vendors at the show. And please, don’t get me wrong: there were some very stellar presentations from analysts (Hello, Merv Adrian) and vendors alike. These observations are from my time manning the booth, giving my presentation, and conducting meetings with customers and prospects. I look for trends, when I am hearing something from more than one attendee. Or, sometimes, simply that moment where I think, “Huh, I never thought of that.” So, here we go:

Tweet #1 – “We’ve been saving unstructured data for a long time and are now looking to do something with it.”

Ok, this one might fall into the “duh” category. However, stick with me. There were 2 types of people who said this to me:

  1. Federal/Government agencies. Yup, a number of them. They all said some derivative of “we’ve been saving unstructured data forever, and we’re just starting to discover what we can do with it.” All of them were very interested in a single cluster, data lake approach to combining their unstructured and structured data. All very cautious, all very concerned with security, and all wanting nothing to do with the cloud. Which brings me to…
  2. Those with a different notion on data location. I had one visitor to the booth come right up and say, “We want to take all our unstructured data and put it in the cloud, and keep all our structured data on-premises.” I must admit, for a moment I was speechless. Not because it is a crazy idea, not because we can’t do it, not because it is a bad business practice. Nope, I was speechless because I had never really considered this approach. I’ve thought of tiered data approaches (like 0-2 years worth of data in memory, 3-5 in some kind of warehouse, and 5+ in MapR, for example). I’ve thought of the cloud for developers’ sandboxes and business data on-premises. You know, all the boring ways to slice this stuff that we do everyday. But this was a new one for me, and he was adamant that his company was going to tackle big data this way. Have I been living too long buried under the Boston snow? Would you ever consider this approach? Let us know your thoughts.

Tweet #2 – “I want to refresh data for my BI and analytics tools on Hadoop when I want to, not schedule them."”

Multiple customers and prospects approached me with this comment. The gist was as one of them so eloquently put it: “I want to sit there and hit refresh on my BI data on Hadoop all day long.” Makes sense to me, but this is way out of my wheelhouse. Enter my good friend Steve Wooledge, VP of Marketing at Arcadia Data!

Check out the on-demand webinar, "4 Ways to Scale Interactive BI and Analytics on a Data Lake" webinar, scheduled for Wednesday, April 5th at 10am PT/1pm ET. At this webinar, you will learn about four different ways to provide analytics on the data lake. What are the pros and cons of each, such as connecting your existing BI tools to Hadoop-based platforms? What are the best deployment strategies for supporting 100s or 1000s of concurrent users across many business units?

And maybe, just maybe, Steve will tell us all how to refresh BI Data on Hadoop all day long!

Tweet #3 – "Hardware? Nodes? New projects for us start in the cloud."

This discussion was a lot of fun and another eye-opener. A prospect sat in on my presentation on “3 Keys to Digital Transformation” and had some follow-up questions. As we were talking, he was more and more convinced that MapR should be a consideration for him. Then I got the golden question: “How soon can you guys spin up a PoC?” Cool, this is going great, I thought. I mentioned that it depended on a number of factors, like node configuration and hardware availability, but we’d need to get the experts out on-site and get the marketing dude (me) out of the way. His priceless response:

“Pfffffft, Hardware? Nodes? New projects for us start in the cloud.” He gave me a bit of this look, too:


Like the comment above on wanting unstructured data on the cloud and structured data on-premises, I think the maturity of this market is leading to strong opinions. This is good. Strong opinions show that big data solutions are being planned for and used in production. Strong opinions on intentions make vendors react. Keep them coming, folks.

This blog post was published March 28, 2017.

50,000+ of the smartest have already joined!

Stay ahead of the bleeding edge...get the best of Big Data in your inbox.

Get our latest posts in your inbox

Subscribe Now