Edge Computing offers a new paradigm for running software applications outside of the Cloud. The Edge is where users and their devices meet the network that connects them. It is defined by its close proximity to end users. It is a platform, like the Cloud, but much more distributed. Sound familiar? If so, read on! If not, read this first:
Your architecture needs an Edge
Introducing the biggest computing paradigm shift in a decade.
A platform without an application is like a hammer without a nail. So what are the killer applications and use cases that make Edge Computing real?
1. IoT Gateways
Edge Computing and the Internet of Things (IoT) go hand-in-hand. With the explosion of new connected devices, everything from your car to your toaster now has an IP address. These new devices are producing a lot of data. So much data that your limited Internet uplink can’t keep up.
Connected devices can consume less backhaul bandwidth by processing the majority of that data at the Edge in an IoT Gateway close to the source, rather than in the Cloud. And should the uplink go down, the IoT gateway can continue to function so you’re not stuck in the dark when your connected light switch and your connected light bulb lose their connections to the Cloud and each other.
Industrial IoT (IIoT) takes these concepts to an industrial scale. Many existing industrial devices expose telemetry that is ignored or lost over time (as much as 97% in fact). By collecting and rationalizing this data at the Edge, industrial IoT gateways can improve the efficiency and efficacy of industrial automation.
IoT and IIoT also serve as a catch-all for several loosely correlated Edge Computing use cases, including connected cars, transportation, and energy. These use cases share the IoT functions of telemetry analytics and localized actuation at the Edge.
2. Network Functions
Routers, switches, and firewalls usually conjure up images of big metal boxes. While network functions have traditionally run on purpose-built appliances, Network Function Virtualization (NFV) has taken these closed systems and turned them into software that can run in a Virtual Machine or even a Docker Container. Unlike the other use cases on this page, network functions must run at the Edge. Packet forwarding and security functions just can’t be outsourced to the Cloud while maintaining real-time performance.
Fortunately, the combination of NFV and Edge Computing make it easier than ever to manage the lifecycle and configuration of these newly virtualized network functions. Network service providers are already considering the shift to NFV as an opportunity for rolling out general-purpose Edge Computing platforms like Multi-access / Mobile Edge Computing (MEC).
Multi-player gaming is bandwidth-intensive and latency-sensitive: a recipe for Edge Computing. By matching gamers based on location and then placing game servers closer to them, multi-player ping latency can hit single-digit milliseconds. The lower the latency between a game console or PC gaming rig and the backend server, the lower the lag. The rise of competitive gaming suggests that the massive gaming community is willing to pay a premium for a better experience.
With Edge Computing, services like Cloud Gaming become viable. Rather than investing in a new PC, PlayStation, or Xbox every year or two, gamers could subscribe to an Edge-hosted gaming service. The Edge hardware is kept up to date and users connect to it remotely. While previous attempts at Cloud Gaming (like OnLive) failed due to issues with real-time latency, the emergence of managed Edge platforms may reignite this model.
4. Content Delivery
The original Edge Computing use case. By caching content — whether it’s a web page, a video, or a piece of music — at the Edge, end-users enjoy a better experience. Edge caching provides a massive improvement over traditional web servers, providing latency on the order of single milliseconds.
This was news 20 years ago; what does it have to do with Edge Computing today? The Content Delivery Network (CDN) market has been dominated by a handful of players, like Akamai and Limelight. They have built out sprawling global cache networks. What if the Edge looked more like the Cloud? Rather than rely on a CDN provider, any content provider, like HBO or Netflix, could spin up their own custom micro-cache at the Edge of the network with greater flexibility and customization than if they had used a general-purpose CDN provider.
5. Machine Learning for Voice and Video Recognition
Voice recognition through Alexa, Google Assistant, and Siri are now mainstream. And, in a world of increasingly connected devices, the number of voice and video capable devices is only going up. Hauling all this voice and video content back to the cloud is costly. Users expect low-latency responsiveness without a taking hit to their data plans.
Edge Computing enables the execution of Machine Learning inference models, such as those used for voice and video analysis, to run closer than ever to end-users and their devices. By converting speech to text at the Edge, a megabyte (MB) voice recording can be converted to just a few bytes of text.
6. Virtual Desktop
While Cloud software is the future of Enterprise applications, many still rely on virtual desktop to get things done. Whether these virtual desktop environments are hosted in a corporate data center or running in the public cloud, they can be slow for remote workers and can even halt productivity under the wrong network conditions.
Edge Computing has the potential to free virtual desktop environments from the data center and enable highly controlled yet localized access for remote workers through Virtual Machine migration. Imagine working on your virtual desktop environment in the office and then having that virtual desktop instance follow you home to your closest Edge site. Like the Gaming use case, lower latency makes all the difference.
7. Video Conferencing
Whether you attend meetings every day or once a quarter, most of us have experienced conferencing software and its occasional shortcomings. Voice delays, bad video quality, and frozen screen-shares are not unusual. These problems can be the result of a slow link back to the cloud, where multiple voice and video streams are multiplexed together.
By placing the server side of voice and video conference software closer to the participants, these kinds of quality problems can be reduced significantly. A globally distributed Edge fabric of voice and video servers enables a much more resilient and responsive user experience for conference participants.
8. Cloud Storage
Everyone has data in the Cloud, but sometimes it can be hard to access. It can be slow or even unavailable. Whether you’re using a service like Dropbox and Google Drive or just trying to access a network file system (NFS) over the Internet, remote storage doesn’t always work the way it should.
Edge Computing can help fix this problem. By placing a storage gateway at the Edge, the gateway can act as a read/write cache. If the Cloud storage service is unavailable, it can offer up the working set of files that are cached locally. On a slow link, it can give the appearance of a fast local storage array, despite an exceptionally slow uplink.
9. Augmented Reality
Finally, the flashy Edge use case. In reality, augmented reality (AR) is an unlikely driver of Edge Computing adoption. While it is possible to offload computation from lightweight AR headsets, the power tradeoffs and current dependencies on specialized hardware make it difficult to justify AR as a use case for a general-purpose Edge Computing platform today. That said, AR certainly fits the mold of the prototypical Edge use case: bandwidth-hungry and latency-sensitive. As this technology advances, augmented reality will no doubt benefit from the augmented computation of the Edge.
No single application alone will drive the emergence of a general-purpose Edge Computing platform. However, taken all together, these use cases will create a platform ecosystem around the Edge, one that can handle heterogeneous workloads while enabling new possibilities not yet imagined. Please feel free to share your own killer use cases for Edge Computing in the responses here. Or connect to me on LinkedIn and let’s start talking.