Gunnar Assmy - Fotolia
The University of Notre Dame sought an enhanced fan experience, one that would let hockey game attendees watch the action from multiple angles on their smartphones. The called for multiple cameras positioned throughout the Compton Family Ice Arena that could deliver real-time video to fans in the stands.
Buying cameras for this endeavor was an expected investment. However, delivering on this vision also required a technology architecture, said J. Nicholas Laneman, co-director of the Wireless Institute in the university's College of Engineering, who worked with the university's Office of Information Technologies to explore this idea.
To create the architecture, the university needed to put more computing power in the devices within the arena itself; in other words, it needed to move to edge computing to get the speed needed to deliver in real time the video feed that each fan requested.
The move to an edge computing architecture required investments in several enabling technologies, including the localized compute power that could come from servers or leased access points from a telecommunications provider, as well as advanced networking and Wi-Fi capabilities in the arena.
"The computing has to be closer to the mobile phone [to make this work]," said Laneman, who is also a professor and associate chair of graduate studies in the electrical engineering department.
Notre Dame officials worked with various vendors on developing a pilot before opting not to move forward, Laneman said -- but its continuing interest in these capabilities is representative of organizational zeal for edge computing.
A place, not a technology
Edge computing is hitting hype status, as many technologists are tagging it as an emerging transformational technology. Yet, edge computing is not a single technology; in fact, one could argue that it's not really a technology at all.
Rather, like cloud, it refers more to a place. Cloud is essentially a virtualized off-site environment for data storage and computing power, while an edge computing architecture puts computation on or close to where the data is generated and used.
Yugal Joshi, vice president of information technology services at the Everest Group, a management consulting and research firm, said his company defines edge computing as "an enabling set of technologies that move data storage, computing and networking closer to the point of data generation/consumption and away from a centralized computing location."
"It's not a device, it's a location, and it typically means the edge of a network," Joshi said. "Edge at the end of the day is an architectural choice."
He and others acknowledge that edge computing has some retro characteristics to it, harkening back to IT before the rise of cloud computing about a decade or so ago and centralized data centers before that, back to an IT era when storage and computing technologies were much more decentralized.
However, experts stressed that an edge computing architecture is much more than decentralized servers. Rather, edge computing is an architecture that powers computation when and where it's needed to deliver actionable information as fast as it's needed. "That's what's making edge today exciting," Joshi said.
Growing organizational interest in edge computing is driving investments in the technology components that enable it, according to researchers and market analysts.
One report, "Edge Computing Market -- Growth, Trends, and Forecast (2019 - 2024)," released in June 2019, stated that the edge computing market is expected to reach nearly $7 billion by 2024, up from $1.2 billion in 2018.
Such figures aren't surprising, considering how much change edge computing will bring to organizational IT infrastructure, experts said. Forrester Research Inc. released a report detailing what to expect with the bold, and somewhat prophetic, title "Edge Computing Will Radically Alter Your Infrastructure Strategy."
However, few organizations are making serious investments yet, said Brian Hopkins, vice president and principal analyst at Forrester. According to Hopkins, Forrester research found that by summer 2019 only 13% of companies have implemented edge computing capabilities and an additional 14% to implement edge computing within the next 12 months.
Experts said the types of organizational edge computing investments will vary depending on the maturity of each enterprise's existing IT infrastructure, as well as the use case that the edge computing will power.
"Understand that the edge is different for everybody; the edge is not just one thing," Hopkins said. "We tell buyers that your investment in edge will depend on the value you want to get."
Some organizations, such as manufacturers that already have made significant investments in industrial internet of things by deploying smart sensors on equipment and upgrading networking equipment, will need fewer investments to make the leap to an edge computing architecture -- perhaps only needing to invest in local computing and analytics capabilities so that the data generated by sensors on the manufacturing floor can be analyzed then and there, rather than moving back and forth to a cloud platform.
Meanwhile, an organization without strong networking capabilities or a cadre of smart endpoint devices will have to buy those components as well as the localized storage and computing capabilities.
Vendors are stepping in to meet demand. HPE, for example, has its Edgeline Converged Edge Systems and OT Link, billed as a "converged OT and enterprise-class IT in a single, ruggedized system that implements data center-level compute and management technology at the edge." Other examples include AWS Snowball Edge, as well as Dell and Schneider Electric's edge computing offerings.
Telecommunication companies are also providing local edge computing offerings. Forrester reported in its Forrester Analytics Global Business Technographics Mobility Survey, 2018, that 27% of the responding global telecom decision-makers are either implementing or expanding edge computing in 2019.
Such offerings will impact enterprise IT opportunities, with Joshi noting that companies will have a "mix and match" of investments in networking, hardware and software to enable their edge computing capabilities.
"Edge will require incremental spend, but on what and how much depends on the use case," he said, adding that CIOs will not build out their edge computing capabilities with only their own IT departments but instead will need "a partner ecosystem."
"It's impossible for one company to do everything. The devices, the networking, the IoT platform, production assets -- there are so many moving parts," he said. "So you need an ecosystem mindset from both an enterprise IT standpoint and from the vendors to make this a success."