MapR's Chief Application Architect Ted Dunning announced seven major big data technology trends and forecasts for 2018.

Innovation in analytics and big data management enables businesses to not only manage the rapid increase in data but also maximize its value. According to IDC, as the volume of data increases exponentially, the share of meaningful data will rise from 22% in 2013 to 37% in 2020.

1. Machine Learning Transforms from a “Technology Trend” to a “Technology Applied to Actual Business”
Machine learning is now expected to be increasingly utilized universally in business. As interest in machine learning grows alongside that in artificial intelligence (AI), its scope of application is expanding, delivering valuable insights to companies across various industries.

The most successful systems are likely to emerge when companies focus more on problems than on tools. It is crucial for companies to have practical goals, scale to access necessary data, and be prepared to ask the right questions for machine learning in order to establish realistic plans to apply machine learning results to actual business operations.

2. 90% of machine learning success is the execution plan (not the algorithm or model).
Effectively managing data is crucial for successfully using machine learning. Data management serves as the foundation for the entire lifecycle, ranging from utilizing input data for machine learning model development to the ongoing maintenance required for operations. Through effective architecture and strategic planning, it will become possible to manage everything at the platform level rather than the application level across various systems based on multiple machine learning tools. In other words, there will be no need to devise a new execution plan for every new project. As such, the importance of efficient machine learning execution planning is steadily increasing. Furthermore, the demand for stream-based architectures and global data fabrics will grow across the entire organization.

3. Laying the foundation for building a multi-cloud environment through the rapid adoption of Kubernetes
While it is predicted that Kubernetes will succeed, the pace of its adoption is already accelerating, so this can be considered the current state of the market rather than a prediction.
Many companies still perceive Kubernetes as a tool for managing and orchestrating computing within the cloud. However, within the next year, Kubernetes will be increasingly used by advanced enterprises seeking to manage and orchestrate computing across all cloud environments, including private and public ones. On-premises computing is evolving very rapidly toward container and orchestration methods.

4. Big Data Systems as a Core Element of an Enterprise (Implementation of Global Data Fabric)
In the past, big data and developed projects were separated, but now big data has become a major corporate asset, and companies are approaching their businesses in a data-centric manner. These changes enable big data systems, encompassing data scale, storage, operations, and analytical access, to become a core element of the enterprise. Businesses will find ways to break down silos and take a comprehensive approach to data through numerous sources to implement a global data fabric capable of computing for true multi-tenant systems.

5. Combine data flows into a data fabric
In 2018, more companies will come to view computing not simply as data residing in databases, but as data flows. Data flows collect key business events and replicate business structures. Integrated data fabrics will serve as the foundation for developing large-scale flow-based systems. Specific fabrics will support various types of computing to enable utilization in multiple contexts. Increasingly, databases will play the role of collaborators and complements to data flows. A new trend in 2018 is to implement data fabrics that provide both unused and in-use data for multi-cloud computing using tools such as Kubernetes.

6. DataOps, a key organizational approach to enhancing agility
Traditional DevOps teams, comprising data scientists and data-driven developers, will transform into DataOps teams. This provides enhanced communication, accelerated goal achievement through collaboration that transcends departmental roles, reduced time to value, and high agility. Managing work in a DataOps format enables the capability to respond to changing environments and take appropriate actions in a timely manner. This enables the flexibility and efficiency to reap the benefits offered by new technologies and architectures.

7. Expansion of IoT Edge
In 2018, not only will companies transitioning data fabric and computing from on-premises to multi-cloud environments, but now, with a maximally scalable data fabric, they will be directly connected to the edge at the location closest to the device or have the data fabric installed on the device.

Ted Dunning, MapR’s Chief Application Architect, stated, “The 2018 technology outlook is based on practical approaches to addressing data-related challenges and insights gained while supporting enterprises in achieving business innovation.” He added, “A recurring issue among companies when adopting new technologies is the desire to implement them efficiently without excessive costs. MapR’s Data Fabric supports customers in achieving innovation while simultaneously reducing the costs of legacy systems.”