As the father of two girls, I have deep appreciation for a fascination with bubbles. They are easy to make, they're shiny, and they float carelessly with the breeze. That said, a big bubble is much harder to create and more fragile than dozens of little ones. The loss is felt more acutely.
Big data is not a big bubble. Despite the curmudgeonly grumbling of my peers at other analyst firms, big data is not going to fail. For one thing, we're more than a decade into the macro-trend, and it's still growing. The investment by industry titans continues apace for analytics software players like AWS, Microsoft, IBM, and SAP. Major hardware vendors are refocusing, as seen with Dell EMC's new push on the Analytics Insight Module, or HPE's infrastructure renewed positioning for big data even as Vertica/Autonomy spin out. Oracle continues to do it all. Big data specialists like Cloudera, MapR, Hortonworks, Databricks, and Qubole all float on rather successfully. The number of startups is difficult to even count. New sparkles like cloud data warehouses, streaming pipelines, machine learning, and AI are all emerging and building new excitement.
And that's the point: Big data can't pop like a big bubble, because it's not one. Big data is many, many clusters of little bubbles of value and functionality. Sure, some will be transient, or undistinguished from their look-alike competitors, and these won't live long, at least not independently. Maybe more of the little bubbles drift up toward the (public) cloud, where they are more readily leveraged together. Yet, overall, the beautiful frothy flow of ideas and innovation will continue apace. The potential of the aggregate big data ecosystem is simply too mesmerizing for businesses to look away.