How the Edge Is Changing Data-First Modernization

From the factory floor to online commerce sites and containers shuttling goods across the global supply chain, the proliferation of data collected at the edge is creating opportunities for real-time insights that elevate decision-making. To reap the benefits, organizations need to modernize with a decentralized data strategy that delivers the speed and flexibility necessary for driving smarter outcomes for the business.

The concept of the edge is not new, but its role in driving data-first business is just now emerging. The advent of distributed workforces, smart devices, and internet-of-things (IoT) applications is creating a deluge of data generated and consumed outside of traditional centralized data warehouses.

At the same time, the availability of 5G connectivity and an influx of robust, cost-effective edge processing power have made it possible to decentralize data storage and real-time analytics processing power and position it closer to the actual data source.

IDC estimates that there will be 55.7 billion connected IoT devices by 2025, generating almost 80 billion zettabytes of data at the edge. In addition, IDC projections show worldwide spending on edge computing reaching $176 billion in 2022, an increase of 14.8% over last year.

A recent survey conducted by IDC and sponsored by Lumen Technologies and Intel Corporation indicates that two-thirds of global IT leaders are implementing edge computing. IDC predicts that by 2023 over half of new enterprise IT infrastructure deployed will be at the edge; by 2024 the number of apps at the edge will balloon by 800%.

Momentum is surging because edge computing opens up a whole new world for data-first business, reducing latency, relieving bandwidth pressures, and enabling fluid data movement. As a result, business users are treated to insights that weren’t possible before, with enhanced agility to act on data in the moment.

“With all this storage and compute capacity at myriad edge locations, we now have the ability to solve new problems that couldn’t be solved before,” says Denis Vilfort, HPE’s director of edge marketing. “The nature of the old centralized data center basically imputed a round trip tax that stopped certain things from being possible at the edge.”

With more compute power at the edge, here is a snapshot of what’s possible:

Manufacturers are collecting data directly from industrial assets such as fluid pumps or oil rigs distributed around the globe or out in the field, driving insights used to optimize processes, identify bottlenecks, keep tabs on quality issues, and even initiate proactive maintenance and take corrective actions in near real time.

• Retailers are tapping sensor data and in-store camera images to identify potential theft across multiple store locations to augment human intervention.

• Autonomous vehicles draw on copious amounts of sensor, camera, and lidar data to power the machine learning and algorithms used to drive the car and react to changing road conditions.

“If we can take the data center with us, we can do all kinds of things that weren’t possible before,” Vilfort explains. “You can’t move these types of applications to the cloud for processing, because you don’t have enough time to process real-time data streams.”

How edge refines data strategy

To support these new data-first business strategies, organizations need to rethink and redefine traditional data warehouses and architectures with an eye toward decentralized strategies that still encompass centralized controls such as governance and built-in security along with frictionless data movement. “We don’t want to apply a centralized paradigm to a decentralized problem,” Vilfort adds. “That’s the promise of edge computing.”

HPE GreenLake brings the benefits of a cloud experience — specifically hardware, software, orchestration, metering, and billing — serving as a unified edge-to-cloud platform that brings end-to-end visibility to a decentralized data estate. The HPE Ezmeral data fabric and file and object store, delivered as a service through HPE GreenLake, integrates files, objects, NoSQL databases, and multiple types of streaming data from existing platforms, including edge locations, into a unified layer to drive analytics applications and promote more intelligent insights.

HPE’s unified analytics capabilities work across a diversity of data types, reaching from the edge to hybrid cloud, in support of decentralized use cases while the machine learning operations (MLops) platform helps automate the end-to-end processes surrounding artificial intelligence (AI) and analytics pipelines, from planning through model development, training, deployment, and monitoring.

Getting edge-to-cloud data strategy right

Turning data-first business from strategy to reality starts with taking stock of where data resides — determining the data footprint — since data is really where the action is. “Being data-first means we get to look at the edge first and then figure out what goes into the data center, the colocation center and so on,” Vilfort explains. “We need to know how much data there is, where it’s going, how long we need to keep it, and who can see it — this is a data conversation and a data management challenge.”

From there, other best practices emerge:

• Heighten the focus on security and governance. As data analytics and AI and machine learning workloads are increasingly directed to the edge, the attack surface for potential security breaches vastly expands. Data-first modernization requires organizations to redefine practices to build in security from the onset, not as an afterthought. They also need to pay close attention to long-standing issues, from data sovereignty to adherence to regional and global regulatory requirements.

• Establish cross-functional teams. It’s important to establish data stewards that hail from different parts of the organization to define and scope data needs as well as to identify all relevant data sources to get a true picture of the decentralized data estate and specifically edge locations.

• Create a center of excellence (CoE). Even though data strategies and data insights are orchestrated from the edge, it’s still important to cross-pollinate ideas and create and promote shared data policies. A CoE can also foster support for strategies, enlist executive buy-in, and help drive the necessary cultural change.

Edge environments promise to open up a whole new world of insights and innovation. Yet with the possibilities comes a paradigm shift requiring organizations to modernize with a decentralized data strategy that propels data-first business.

Click here for more information.