LAS VEGAS, NV – Edge computing strategies are crucial for an organization to meet its edge goals, Gartner Analyst Thomas Bittman said during the firm's IT Infrastructure, Operations & Cloud Strategies Conference. Gartner predicts that by 2025, 50% of edge computing solutions deployed at enterprises before a cohesive strategy has taken shape will fail to meet deployment time, functionality, and cost goals.

"Digital transformation is absolutely happening, and it doesn't just happen to the backend. That's what the cloud is giving us. It happens on the front end, and it's going to happen at the edge," Bittman said. He defined edge computing as part of distributed computing architecture "anywhere between that backend cloud or data center and that physical edge. Everything is fair game."

According to Gartner, key considerations of an edge computing strategy include a clear vision, specific use cases, potential challenges, standards or best practices, and deployment execution.

"You have got to start with a vision," and it should integrate the edge strategy with other areas of the organization like cloud strategy, Bittman said. "This isn't something that stands alone." He recommended organizations draft a vision statement for edge that ties tech investments to business value and disseminate that statement to gather executive buy-in and sponsorship.

All edge use cases are unique depending on the type of industry, equipment, and interactions in question. This high diversity makes it incredibly difficult to put a general edge plan or architecture in place, and it makes it difficult for vendors in the space to turn a profit, Bitttman explained. "But that's going to change."

Identifying and managing edge computing risks, inhibitors, and challenges is another important piece of designing an edge compute strategy. Bittman recommended exploring how challenges like those associated with distributed computing, distributed data, security and privacy, and extensibility may impact an organization's edge ambitions. Enterprises should decide which challenges will need to be overcome and put a special focus on solving those challenges in terms of management, investment, and skills.

Bittman also highlighted how the lack of edge computing standards and broadly accepted architectures will render more than 85% of enterprise tech stacks incompatible in reality. He suggested organizations create an architecture function focused on edge computing, design and maintain architecture standards, determine best practices outside the enterprise and from enterprise deployments, and develop specific edge computing skills, roles, and responsibilities within the organizational structure.

In addition, edge proofs of concept (PoCs) shouldn't be blindly trusted, Bittman argued, citing Gartner research that predicts 50% of edge computing solutions that worked fine as PoCs will fail to scale in production through 2022. To that point, he recommends enterprises require highly specific PoCs that address unique edge requirements like scale and tolerance to disconnection.

And then once an edge solution is deployed, Bittman noted requirements will likely change, the scale will grow, and new use cases and risks will emerge over time. Enterprises should design a process to manage the evolution of edge in production, in particular relating to how workloads, data, scale, and use cases shift over time.