Making your edge boost your bottom line
Edge computing promises to deliver the next level of improved operations, unlock new opportunities and increase business agility—not just in a single plant or location but across all locations. Every extra dollar that can be generated in one location is multiplied by the number of locations a given business operates. Making an extra $1,000 per location across a thousand locations delivers $1 million to the bottom line. It is this multiplication effect that makes edge computing so attractive and worth pursuing.
Please read: Your edge, your future
Moving computing closer to the data
By pushing compute power to the distributed edge locations of an enterprise, data at the location can be analyzed in real time. This is something cloud computing cannot do. Moving data to and from the cloud simply takes too long, and doing so pushes the time needed to analyze data and take corrective action outside the window of opportunity. Quick decisions are better made at the location where the activity takes place.
Edge projects can be complicated, however. A comprehensive edge computing solution may have hundreds of cameras and sensors as input devices. In addition, a given configuration may be replicated within a given site and across sites in different geographies.
For example, a factory producing 5,000 complex machines per month per production line has only 376 seconds to manufacture each unit. Burdening the production with a cloud-based QA process can add up to 79 seconds to the build time. Eliminating this cloud round-trip tax per unit increases the factory capacity by 21 percent.
The plant manager cannot ignore this ability to create "free capacity" and must therefore adopt systems that maximize the output of production equipment. To unlock the extra upside, QA data analytics must take place in the factory itself.
Edge computing is the answer, as it is much faster than cloud computing. And as an increasing amount of data is generated in the factory—at the edge—this strategy becomes increasingly attractive over time.
Please read: What's the edge really about? We asked our experts
Here are three components that can contribute to a successful edge computing strategy:
All as one
Edge computing cases vary widely in style and scope, ranging from a shoebox-size data center on an oil rig to thousands of Bluetooth low-energy tags at a retail site. So organizations need to choose the technology mix that best applies to the particular job. Edge computing installations can be stationary or highly mobile. And, as application needs change, enterprises have to be able to move workloads where they're needed most, so they can facilitate the business outcome they want to achieve.
Managing edge computing installations across the globe is a large undertaking, one more demanding in scope than managing a data center in a single location. Adopting a cloud-anywhere strategy, delivered in an as-a-service model, can help enterprises reduce complexity, retain control, and scale edge projects to the point where they can generate positive business outcomes.
Moving real-time computing, analytics, and control to each edge location does not mean complete autonomy for each location. While the location may operate autonomously, managerial oversight is still needed across locations. Orchestrating edge locations as a pool of distributed assets borrows from the trucking industry's concept of fleet management, whereby the organization inserts a control point above the locations that enables the entire distributed enterprise to move as one.
This is where the centralized nature of cloud computing provides a great opportunity to aggregate common data across sites and provide a single version of the truth across operations. Workloads can be published from a central location to each site as older workloads are retired, packaged in a container or a virtual machine, or offered as a service. What's important is moving fast and retaining flexibility.
If a coffee shop chain wants to improve customer experience as well as profitability, adding compute power at its edge—at each of its 3,000 shops—will help. The profits can quickly add up by generating an additional 25 cents per cup in revenue. But rolling out new technologies across locations comes with cost and risk. The management challenge: How does the business gain the benefits of edge technologies while balancing investment capacity and risk? With an as-a-service approach, the company can structure a plan to pay a service organization to manage the rollout as profits roll in.
As-a-service models are maturing, and they're going to be important for edge computing. Edge projects need to scale quickly from proof of concept to production, and this requires economic flexibility. Outcomes can be projected during a pilot program, and project payouts can be structured based on money saved or revenue generated.
Large scope and scale have a direct impact on the economic performance of an edge-driven company. But managing vast networks that are geographically dispersed and technologically diverse can be challenging. Just because you have a great idea for an edge project doesn't mean you have the expertise to envision the parameters, deploy the hardware, scale the operations, and optimize the entire system. It's far harder to fix a series of camera glitches across multiple locations than to fix a server malfunction in a data center down the hall. What do you do if something goes wrong? The answer is straightforward: partner with a service provider with expertise, technology, and global reach to solve problems proactively.
The path forward
As edge strategies evolve, companies will face challenges. Taking advantage of flexible technologies, payment plans, and management models, they can clear initial hurdles and scale their projects to deliver the outcomes they desire.
By combining a global partnership with an as-a-service model, the benefits of edge computing are now within reach. Any business or government entity can improve operations by moving to an edge-driven strategy. And doing so is now more achievable than ever.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.