As 2014 nears its end, it’s the perfect time for our second annual Technology Predictions series in which industry experts share their predictions about the hot tech trends that they think will take center stage in 2015. We’ll be sharing all of these predictions with you over the next couple of weeks.

This next set of 2015 Technology Predictions on Big Data comes from Phu Hoang, CEO and co-founder of DataTorrent, Inc., a provider of an enterprise-grade, real-time stream processing platform on Hadoop.

.

.

If 2014 was the year Big Data went mainstream, what should we expect in 2015? Given that the amount of human and machine-generated data related to Internet of Things (IoT)-connected devices has exploded over the past year (Gartner, Inc. expects the number of IoT-connected devices to grow to 26 billion units by 2020), enterprises will be forced to reevaluate their Big Data tools to effectively use their data and take action in business time. In 2015, look out for the following:

.

Prediction No. 1: Taking action in business time: 2015 will be the year of “fast Big Data”

Data is not only big but it is also fast. Data is rapidly being generated from mobile devices, factory sensors, wearables, retail transactions, online advertisements and many other sources. The insights from that fast, Big Data disappear as rapidly as they appear so taking action in “business time” is critical. Business time is the speed at which a business must service its customers, and it ranges from sub-seconds (real-time) to minutes and hours.

.

Prediction No. 2: Not your grandfather’s Hadoop: Hadoop on the edge

2015 will be the year that Hadoop makes the move to the edge. Traditionally, Hadoop was positioned as cheap storage and was meant to live in a data warehouse. We are starting to see Hadoop moving out to the edge–out of the data warehouse. Edge doesn’t always mean “embedded,” and many edge locations need scale out and fault-tolerant compute and storage in small form factors. Hadoop is a very natural fit.

.

Prediction No. 3: Making it consumable: Ingestion is the killer app

As more connected devices come online, organizations are realizing that Big Data is actually a lot of little data and that it’s only growing in size. With this exponential growth in the amount of data, organizations need to be able to ingest that data in real time in order to maximize the business value of that data.

.

Prediction No. 3: Deploying quickly and inexpensively: Streaming analytics with Hadoop-as-a-Service

Standing up and running your own Hadoop cluster takes considerable time and investment. The Hadoop-as-a-Service offering is gaining traction and will continue to proliferate through 2015. Driven by customer demand, we see significant vendor interest in offering streaming as an integral capability of the Hadoop-as-a-Service platform. Customers want to gain insight while the data is ingested in the environments of their Hadoop-as-a-Service vendor.

.

Prediction No. 4: Explosion of new services: The ubiquity of YARN

YARN, which takes Hadoop beyond batch to enable broader data processing, has considerably matured throughout 2014. YARN is running in production across the larger enterprise and allowed for expansion onto Hadoop. As we head into 2015, we will see Hadoop become a rich operating system that will begin to offer delivering services.

.

Prediction No. 4: It’s not just for geeks: Look for the democratization of Big Data in 2015

As more businesses join the Big Data party in 2015, organizations will require Big Data that speaks to more than just the Data Scientist. In 2015, be sure to look for clear, easy-to-use dashboards on analytics, offering more self-service analytics capabilities that do not require data to be moved out of the Hadoop cluster.

.

2015 Technology Predictions graphic courtesy of SiliconANGLE