Edge monitoring and management in the age of IoT
Modern enterprises have experienced a deluge of data over the past few years. Now, with billions of IoT devices connecting from every nook and cranny of the planet, and business units transforming through technologies such as edge AI, enterprise IT finds itself deploying and supporting infrastructure in a decentralized operating model.
It makes sense when you think about it. In this age of the customer, no organization can afford the kind of latency, security, and connection issues that can occur when transmitting massive digital signals between core data centers and far-off places. So locating micro data centers closer to the 56 billion IoT devices that are expected to be operational by 2025 is a logistical no-brainer. What's more, it's a money saver: Analysys Mason's Gorkem Yigit says this move could save up to 20 percent on operational costs. Not surprisingly, because of these advantages, edge computing is exploding, with 70 percent of enterprises expected to be running varying levels of data processing at the edge by next year. The worldwide market for edge computing is projected to reach more than $250 billion by 2024, according to IDC.
The edge is clearly hot, but here's the rub: It's new, and deployments often go sideways. That's why it's vital to not only have an edge strategy but to have one built on modern-day best practices.
"Every analyst, every consultant, every adviser out there says, 'Have a plan for the edge.' But as Mike Tyson says, 'Everyone has a plan until they get punched in the mouth,'" quips Matt Kimball, a senior data center analyst at Moor Insights & Strategy. "That is so true as it relates to edge. Everybody has a plan, but too often they're written by smart IT folks that have never done an edge computing deployment, and so they struggle because they aren't fully prepared for the challenges to come."
Managing edge complexity
One of those challenges is having IT leaders who earned their stripes in traditional data centers trying to apply old-fashioned enterprise tools and processes to edge environments. It rarely works, analysts say, because the edge is an entirely decentralized world, complete with its own set of inconsistent capabilities and requirements.
"It's one thing to manage a thousand servers in a single building, but when you have to manage thousands of servers across thousands of different locations, it becomes a completely different story," says Dave McCarthy, research vice president in the infrastructure practice at IDC. "I was surprised to learn that many IT leaders feel they need training for the edge because everything is so different from what they've been doing everywhere else, so complex, that they struggle to know how to manage it."
Please read: Your edge. Your future.
Complexity is the operative word. By next year, Gartner says more than half of enterprise-generated data will be created outside the data center or cloud, up from less than 10 percent in 2019. Edge servers will pop up everywhere—in factories, the back of department stores, sports arenas, hospital basements, electrical substations, you name it. The more that happens and the farther those servers move from core IT resources, the more manageability and security risk gets introduced.
Old solutions mistakenly applied
In early edge computing deployments, many IT staffs tried addressing that risk with legacy network management tools. But executives say tools such as Chef, Puppet, and Ansible proved ineffective at seeing, integrating, and handling the dramatic increase in edge workloads and data volumes coming their way.
"Network connectivity becomes the Achilles' heel of edge deployments," says Aaron Carman, chief technologist for North America at Hewlett Packard Enterprise. "For many organizations, the whole network piece was an afterthought in edge computing deployments. IT leaders don't quite appreciate the uniqueness of network connectivity for all those edge sites. They don't account for how bandwidth, latency, and connection capabilities are vastly different. They also try using enterprise data center tools for edge infrastructure deployments to get a handle on all their previous edge environments and find out too late that it's never going to work."
Some organizations also try to use some of the same enterprise servers they typically deploy in data centers at the edge but ultimately find that they are not rugged enough for the job. "Understanding site capabilities is key, which is why some vendors have been building out server portfolios that are specifically designed for edge deployments," says Carman.
Security issues still arise
Physical and digital security has posed another challenge. McCarthy points out that although the edge can be more secure than the data center or a cloud computing environment, its distributed nature poses distinct vulnerabilities.
Remote sites such as retail stores, for instance, typically have lighter security than a company's headquarters, he notes. The problem: As you add remote sites, you widen the potential attack surface.
"Anytime you put equipment in some remote place, you're increasing the possibility someone could tamper or walk off with it," McCarthy says. "There are all sorts of issues that could arise, such as someone plugging USB keys into the equipment [to corrupt or steal data]. Traditional data centers were designed to support a highly secure operating model. When you get to the edge, it's anything but that."
McCarthy adds that the more remote organizations implement edge computing, the more difficult it becomes to oversee and enforce identity, authorized access, and other security policies. The local staff isn't typically there to help, and few IT departments can monitor and manage everything centrally using their existing tools.
Best practices to the rescue
While noting significant challenges with edge computing, analysts such as McCarthy and Kimball are optimistic about edge computing's future. The technology behind it is evolving, they say, and organizations are increasingly applying modern best practices to overcome common hurdles.
For example, tech vendors such as Red Hat, SUSE, and VMware are retooling existing product lines to provide better management capabilities for smaller, remote configurations, says McCarthy. "They're making it possible to purposely treat the edge as part of a consistent infrastructure, which should be a goal of all these deployments," he says.
More organizations are also employing third parties or software-as-a-service offerings for managing and securing edge deployments rather than trying to control their continually expanding digital ecosystems themselves. In fact, IDC predicts that by 2024, more than 75 percent of infrastructure in edge locations will be consumed or operated via an as-a-service model, as will more than half of data center infrastructure.
Please read: Manage your edge data before it becomes a problem
Using third-party consultants or managed service offerings saves organizations time, energy, and money while ensuring their operations run as smoothly and securely as possible. At the same time, outside experts can help organizations keep systems updated, no matter where they sit. They can also help enforce strong identity and access management policies, including zero trust, where anyone accessing a network has to be completely and continually authenticated.
Tech vendors are focusing more on the physical security side of edge computing as well. For example, some servers now come with a tamper-proof chassis, accelerometers that sense when a piece of equipment is being moved, and even GPS features that can locate a server should it vanish.
Carman says autonomous edge is a future management capability leveraging intelligent technologies to self-manage edge deployments that suffer from connectivity issues or lack support personnel. AIOps edge platforms leverage artificial intelligence and machine learning to automate edge device management processes. That's important, because in dealing with edge environments, traditional models rarely have the bandwidth necessary to manage the edge directly.
"Organizations need an edge strategy for managing and securing edge workloads and data because the speed at which business is expanding is outpacing IT's ability to handle it," he says. "Edge is bringing a whirlwind of opportunity and challenges our way, and we all need to be ready to enable it for our business."
"For many organizations, understanding connectivity and network characteristics at scale is an afterthought in edge computing deployments."
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.