The Razor’s Edge: Taking A Clear View As Compute Moves To IoT Endpoints and Systems Scale
November 6, 2019
Occam’s razor is based on a philosophy for solving problems: “Entities should not be multiplied without necessity.” This straightforward notion is attributed to English Franciscan friar William of Ockham (1287–1347).
Simpler is better, and according to Occam’s razor, the simplest solution is most often the right one.
800 years ago, the followers of the Occam’s razor philosophy could not have imagined a world of hyperconnected things, fusing the physical and digital realms, but the leaning toward simplicity may be more relevant than ever. Given the variety of solutions possible when it comes to managing and optimizing the IoT edge, the complexity of the alternative ways to treat connectivity, data processing, data storage, data sharing, and the evolution of applications which is where the real value resides can feel nearly incomprehensible.
There are dozens of ways to treat the edge, with so many combinations of endpoints, actuators, gateways, transmission networks both local and long-distance, clouds, local and distributed applications and more.
I’ve studied and continue to study various technical and business approaches to developing, managing and scaling edge solutions for IoT and the Industrial IoT and what I’ve learned in the process is that clarity is sanity.
My definition of the edge comes down to this: the edge is where you wish to process data that you don’t want or need to send to the cloud or to your corporate IT. While cloud computing is, of course, transformational – the cloud simply isn’t the right solution for all use cases.
If your solution requires the collection of data from a lot of devices, machines, and sensors, and needs to process it before it moves to the cloud, if it does move to the cloud, your solution can be dramatically improved with edge compute, either embedded using microprocessors in the end-points themselves, or through local communications with gateways that have a short “round trip” and thus support near real-time or even real-time applications requiring low latency and, in some cases ultra-low latency data transmission.
Computing closer to the edge or at the edge not only improves the performance of IoT systems, but dramatically reduces the costs of cloud, improving speed and security, and durability. If the cloud goes down, mission-critical IoT solutions need to keep functioning. If you are running smart factories, if you are protecting a train crossing, if you are securing a building – you need the system to always be available and operating in such a way that automation is successful, and advanced applications including AI contribute.
Take augmented driving and the full range to autonomous driving; with vehicles on the roads and people’s lives at stake, there can be no “hoping” that the signaling and data processing between a vehicle and the cloud will work. New mobility, which is increasingly being bundled with massive smart city infrastructure rollouts, needs data to be processed at the source, while still feeding cloud-based systems for oversight but not on-the-ground operations.
Many say the edge is going to grow beyond cloud going forward, and we have seen similar phenomena in the history of technology and networking over the last twenty or thirty years. Our fixed-line phones are being abandoned, our mobile devices are becoming more powerful and necessary, our smartphones are replacing PCs and laptops for many (as the evolution from mainframes to PCs and then connected PCs paved the way for our understanding of constantly evolving computing paradigms).
Billions of dollars of investment are going into developing edge solutions – and supporting them.
This includes rollouts around the world of 5G networks and investments in edge data centers.
Look for the edge to start to segment itself, with very clear definitions of physical edge management, digital infrastructure management, and application management. This is where we will start to see massive value being created when we are all clear on the way we can develop systems that are simply more secure, more efficient, faster and more resilient than we’ve ever needed before.
We’ll see different and clearer definitions of what edge means for fleet management of mobile assets vs. fixed asset management, for example, equipment in restaurants which are instrumented for applications including energy savings, predictive maintenance, yield analyses, compliance with food safety regulations, and more.
Because edge computing is relatively new, and such a potentially widespread industry, being vague is benefitting organizations when they define what edge means to them.
Beware of the BS: for example, one edge computing company offering Software-as-a-Service and Platform-as-a-Service uses the phrase “edge computing” when they only compute in the cloud.
Another IoT device company says they support compute at the edge, when in fact their devices are connected to the cloud (edge) or the wireless network (edge). This kind of marketing language blurs the definitions and confuses buyers.
ClearBlade, an edge computing software company enabling enterprises to rapidly engineer and run secure, real-time, scalable IoT applications, doesn’t just run on one of these definitions of edge.
Let’s also be clear to not confuse embedded firmware with edge computing. Firmware has been around for decades and is hardcoded for specific devices and low-level solutions. Edge computing in our vernacular and practice is flexible, enabling entire applications to be deployed to multiple edges via a centralized location.
ClearBlade runs everywhere (in the cloud, on the edge of the cloud (like in data centers and hardware), at the edge (on an IoT gateway), and on-premise (on a server behind a corporate firewall). That’s table stakes going forward. What differentiates us, however, is that we are developed and are advancing the only all-edge, fully customizable, adaptable IoT edge products, able to be integrated with existing and legacy systems without altering the architecture.
No matter what technological, product, or industry changes may appear in the future, users of ClearBlade know that they won’t have to rework their whole platform– because ClearBlade will adapt with them. No matter what their protocol or business model is, ClearBlade has a platform to match it, which is what gives us our CLEAR edge.