Edge computing lessons learned

In a world in which even small amounts of latency are unacceptable, IT leaders must build an IT infrastructure that delivers a near-real-time experience to employees and customers.  To meet that challenge, many are turning to edge computing architectures. Edge locations might be in factories, stores, or branch offices, depending on the company and industry.

Putting hardware, software, and network technology at the edge, where data originates, can speed responsiveness, enable compute-hungry AI processing, and greatly improve both employee and customer experience. But edge computing is unexplored territory for many and requires ongoing evaluation of performance and cost-effectiveness.

Despite challenges, the move to the edge is gathering steam. According to IDC research, edge computing is growing worldwide at 15% annually and will reach $232 billion in 2024. Further, 74% of companies plan to increase their edge spending this year. AI is a significant driver. “Edge computing will play a pivotal role in the deployment of AI applications,” says Dave McCarty, research vice president for cloud and edge services at IDC.

Casey’s, a U.S. convenience store chain, is relying on edge architecture to underpin the company’s forays into AI. One application enhances the engagement of customers with Casey’s loyalty program by offering personalized purchase suggestions, fuel discounts, and rewards points.

Moving compute resources to a remote location is seldom an easy task. Successful edge implementations depend on multiple factors, including team building, technology choices, and partner selection. Rather than build out the edge infrastructure themselves, some IT leaders opt to use the edge computing service of a cloud provider. Although edge implementations can cost more, eliminating latency can be worth the expense. To that end, IT leaders should perform a careful analysis of ROI before, during, and after an edge implementation.

“Understanding the many implementation approaches and as-a-service options, as well as the ability to measure return on investment, is critical to building an effective edge strategy,” asserts Jennifer Cooke, research director for IDC’s Edge Trends and Strategies.

Edge architectures vary widely. Some organizations might put hyperconverged infrastructure (HCI) appliances at edge locations, connected over 5G networks, high-bandwidth landlines, or satellite links to a central site. Even when significant technology investments are made at the edge, the central site usually retains an important role in managing and monitoring the edge infrastructure. A central location might also be the nexus of data storage and backup.

Flexibility is the key

Just as there is no single paradigm for a successful edge implementation, edge architectures must evolve over time, sometimes increasing the edge commitment and sometimes decreasing it. This IDC PeerScape report tracks the experience of three successful edge implementations and finds that the key to success over time is flexibility: the ability to grow and change as needs change.

For example, Casey’s finds itself in an industry where growth through acquisitions is a fact of life. Since 60% of the convenience store industry consists of individual stores and chains with 10 or fewer stores, many of these smaller players get acquired by larger chains. Because of this, Casey’s CIO Sanjeev Satturu deployed a flexible architecture to enable the assimilation of acquired companies. “We wanted a virtualized environment that would be technology-agnostic so that we could integrate with any software or hardware. We didn’t want vendor lock-in,” he explains.  

Other factors that led to the edge architecture also can change. For example, thin network bandwidth might have driven the edge implementation to begin with, but an influx of bandwidth might mean the edge architecture is no longer cost-effective, leading to some applications being moved back to the central site. For example, a consumer services company established an edge architecture for a variety of its far-flung locations, in part because of poor bandwidth. But at one location, a significant boost in the available fiber-optic bandwidth led the company to recentralize applications and infrastructure.

Re-assess, re-evaluate, optimize

Ensuring that an edge implementation is optimized for an organization’s needs requires ongoing ROI assessments. “Our architecture team is analyzing all of our applications and whether they should run in the datacenter, in the cloud, or at the edge. We have a list of applications we’re looking at rewriting,” says the VP of IT at a financial technology company, which moved a time-and-attendance application to the edge so that its clients would not suffer excessive latency.

The company is now assessing whether any more of its 438 applications should also run at the edge. Tops among the contenders is an electronic payment application that would speed the completion of transactions, enabling the company to receive compensation from clients faster. However, the executive suggested the financial technology company could eliminate its datacenter altogether, opting to run applications on either a traditional cloud-based service or on an edge-based service such as that of CloudFlare.

Although the company might increase the number of its edge applications, it is very unlikely it will rely entirely on edge implementations. “Some people are saying everything will go to the edge. But we’re not seeing that. The edge is not for general application delivery. There are certain characteristics an application must have. For example, the applications must benefit from increased continuity and elimination of downtime,” the VP explains. Satturu of Casey’s agrees: “Edge is an important part of our technology foundation that will help enable better outcomes. It’s not a silver bullet but must be combined with good application, security, and data design to unlock its full potential.”

Learn more about IDC’s research for technology leaders.

International Data Corporation (IDC) is the premier global provider of market intelligence, advisory services, and events for the technology markets. IDC is a wholly owned subsidiary of International Data Group (IDG Inc.), the world’s leading tech media, data, and marketing services company. Recently voted Analyst Firm of the Year for the third consecutive time, IDC’s Technology Leader Solutions provide you with expert guidance backed by our industry-leading research and advisory services, robust leadership and development programs, and best-in-class benchmarking and sourcing intelligence data from the industry’s most experienced advisors. Contact us today to learn more.

Stanley B. Gibson is an adjunct research advisor with IDC’s IT Executive Programs (IEP), focusing on digital transformation, IT leadership, IoT, cybersecurity, and data management. An award-winning technology journalist and an experienced speaker at worldwide industry events, Gibson has been executive editor at eWEEK and PC Week, news editor at Communications Week, and software and systems editor at Computerworld. Gibson has interviewed the IT industry’s leading executives, including Oracle CEO Larry Ellison, Dell CEO Michael Dell, IBM CEO Samuel Palmisano, Cisco CEO John Chambers, Microsoft CEO Steve Ballmer, and many others.