Understand why edge computing technology matters

To deliver on IoT's promise of a digital revolution, connected devices need access to localized compute, storage and networking resources. Welcome to life on the edge.

When it comes to the potential to break the internet, Kim Kardashian has nothing on IoT. Between ubiquitous phone and tablet use and a burgeoning population of connected things, a staggering amount of traffic threatens to overwhelm traditional centralized networks.

Edge computing technology -- also known as fog computing -- promises to address the imminent data deluge by offering a distributed IT architecture that moves data center resources toward the network periphery. By allowing processing and storage to occur at or near edge devices -- rather than at a centralized, geographically distant data center -- an edge computing topology can reduce latency for time-sensitive applications, support IoT performance in low-bandwidth environments and ease overall network congestion.

Why do we need edge computing technology?

  • Latency: Gartner predicts that by 2021, 25 billion connected things will fill our professional, personal and public spaces. In addition to generating unprecedented amounts of raw data, many IoT applications will demand next-level responsiveness and reliability. Consider intelligent traffic signals in an autonomous vehicle environment, for example, or safety sensors at a manufacturing plant -- latency or network outages could cause accidents and cost lives.
    By virtue of physical proximity, time-to-action drops when data analysis occurs locally rather than at a remote data center or cloud. With the benefit of edge computing technology, IoT and mobile endpoints can react to critical information in near real time.
  • Congestion: Edge computing technology can also ease growing pressure on the wide area network, improving efficiency and keeping bandwidth requirements in check -- a significant challenge in the age of mobile computing and IoT.
    Consider, for example, that a single industrial sensor might generate multiple data points per second, with tens of thousands of sensors in a manufacturing plant churning out millions or even billions of updates per day -- few of them actionable. Rather than overwhelming the network with a constant flood of relatively insignificant raw data, edge computing devices can analyze, filter and compress it locally -- periodically sending relevant, aggregated data onward to the cloud or an in-house data center for long-term storage and big data analysis.
  • Bandwidth: An edge computing topology can also support IoT devices in environments with unreliable network connectivity, such as cruise ships, offshore oil platforms, rural agricultural plants, remote military outposts and ecological research sites. Local compute and storage resources can enable continuous operation even with a hit-or-miss connection to the cloud.
Edge computing infographic

Types of edge computing technology

In some instances, limited edge computing capabilities are embedded directly into connected devices themselves -- such as a wearable health monitor that analyzes heart rate data or a smart security camera with built-in facial recognition processing capabilities. As computer component costs drop, in-device edge computing technology becomes more feasible.

An edge computing topology can reduce latency for time-sensitive applications, support IoT performance in low-bandwidth environments and ease overall network congestion.

Edge computing can also occur in stand-alone network devices, such as routers, sensor hubs or IoT gateways. More intensive compute and storage resources reside in localized edge servers or in small clusters of edge servers that form micro-data centers, which mobile carriers are already deploying at their 5G cellular base stations, according to Gartner. Some IT pros also consider larger, regional data centers to be under the broad umbrella of edge computing.

Emerging use cases span industries, with applications in energy, finance, healthcare, manufacturing, retail, transportation, virtual reality and more.

Challenges of edge computing technology

Edge computing technology is not without its challenges, of course. As Patrick Hubbard -- head geek and senior technical product marketing manager at SolarWinds -- points out, the more intelligent an edge device, the more intensive its configuration, deployment and maintenance requirements.

Organizations will need to decide on a case-by-case basis if distributed computing benefits like cheaper WAN connectivity justify the increased overhead at the network's periphery. Similarly, Gartner research director Santhosh Rao cautions that the costs associated with deploying and operating edge computing technology can pile up quickly. IT leaders should make sure a project's benefits outweigh its costs.

Security is also a major concern, with some IT pros worrying that a decentralized computing architecture makes a network more vulnerable to attack by creating a plethora of backdoor entry points. Others argue that placing an edge-computing gateway between network endpoints and the internet can actually improve security, with more data processed and stored locally and less traveling to and from the cloud.

Despite such uncertainties, analysts expect organizations will increasingly rely on edge computing technology in the years to come. According to Rao, just 10% of enterprise data was created and processed outside of a centralized data center or cloud in 2018. He predicts that number will climb to 75% by 2025.

This was last published in April 2019

Dig Deeper on Edge computing