When looking for answers around edge computing, why not reach out to an industry expert?  VMblog sat down and spoke to Andrew Foster, product director at IOTech, to learn more about edge and some of the key challenges of deploying at the edge.

VMblog:  Why do you believe that edge computing is key to industrial transformation?

Andrew Foster:  Most companies are seeking to improve the efficiency of their operations. The way that this is being achieved is to first build a digital picture of their industrial processes. From that picture organizations can use the real-time information to optimize existing systems in order to improve production output or deploy new use cases such as a predictive analytics application to minimize equipment downtime. New open edge computing solutions, such as IOTech's edge platform software, play a pivotal role in this digital transformation by enabling connectivity to the OT equipment and unlocking access to the valuable machine data.

VMblog:  Are edge systems providing value today?

Foster:  Yes. Edge computing solutions are being used to normalize, aggregate and transform the data from multiple data sources and support local data processing at the edge for more responsive analysis and decision making. They are also used to enable interoperability and convergence between the previously siloed OT systems and backend IT/cloud systems.  Edge computing is enabling the dedicated traditional hardware-centric OT world to become "software defined" by supporting modern computing technologies. The result is that edge computing is enabling a new generation of industrial systems that are more open, flexible and extensible than previously possible. This allows new use cases to be deployed more quickly and systems to be upgraded without having to replace existing investments.            

VMblog:  Is edge computing going to replace cloud computing?

Foster:  As part of an organization's digital transformation, there is a clear acceleration from fully centralized cloud-based systems to distributed architectures driven by edge computing. Cloud computing alone is unable to handle the vast amounts of data created by the huge number of connected devices predicted. It will not meet the need for local insights from the latency-sensitive applications on which they depend.

What is also clear is that most systems will not rely 100 percent on edge computing solutions, either. In fact, fully autonomous edge applications are quite rare. Most systems require a hybrid solution consisting of both edge and cloud components.  

To be successful, users require the ability to utilize cloud resources for heavy-duty applications, while using edge computing for lighter-weight processing and local, real-time insights.

As OT and IT systems converge it is critical to achieve seamless interoperability between edge and cloud environments, which have traditionally been viewed as two separate worlds. Open edge software platform technology, such as IOTech Edge Xpert and Edge Xrt,  is playing a key role by providing the OT and cloud connectivity and edge processing capabilities needed to acquire and normalise OT data, process and transform the data, and then send it seamlessly to the cloud for further analysis or storage.    

VMblog:  Where are current edge computing solutions lacking and what are the key challenges of deploying at the edge?

Foster:  Companies want edge solutions that are easily installed and even easier to own and operate.  They also want solutions that are already working at scale and immediately demonstratable. This is a challenge for solution providers because no edge/Industrial IoT solution currently does it all (and be skeptical of any company that says they do).

Edge elements must be fully integrated into a user's choice of technology, including hardware, sensors, devices, network, cloud providers, data visualization, analytics and security.

Edge solution providers such as IOTech are investing heavily in new product development, as well as partnering where it makes sense to integrate complementary technologies. The goal is to offer production-ready solutions for organizations wanting visible ROI today, not next month or next year.

As companies move toward full scale deployments, a specific area that has not been fully addressed is how to manage an edge deployment at scale. This includes being able to provision, manage, monitor and update not just the edge applications, but also the edge devices that they run on throughout the lifetime of a system, which could be in production for many years.

Application orchestration mechanisms such as Kubernetes don't address the full scope of the problem. As a solution originally developed for cloud/IT environments, Kubernetes et al., in many cases can't support the specific requirements and constraints of deploying at the edge. IOTech has created an open edge management solution called Edge Builder that has been specifically designed to address the needs of managing an edge deployment at scale.  

VMblog:  Are there specific industries that are benefiting from edge computing more than others and which markets are seeing the fastest adoption?

Foster:  We are seeing take-up across a broad range of industries. However, there are a number of device/data rich industrial verticals where IOTech is seeing significant growth.

Probably the largest market for the adoption of this new generation of edge computing solutions is industrial manufacturing and automation. Acquiring real-time data from a plant and creating a digital picture of a manufacturing process is the first step to optimizing an organization's operations. Edge computing is playing a crucial role in aggregating and processing the data from a wide range of disparate data sources, including video, and is key to supporting a whole new range of data driven use cases, such a predictive maintenance or production line fault detection.        

Building automation is another area where IOTech is involved in a large number of projects. A building environment incorporates many different individual systems, including HVAC, lighting, and access control. Edge computing is enabling the aggregation or "fusion" of real-time data from these siloed systems to drive local analytics at the edge that can be used to automatically control and optimize the building. For instance, to automatically reduce the temperature of a room if unoccupied in order to reduce energy consumption, and to adjust the environment for the comfort of the building users.

The third major area where we are seeing large scale use of edge computing is in Smart Energy. For example, IOTech is partnering with a number of the largest Battery Energy Storage Systems (BESS) OEMs where edge computing solutions are enabling the exploitation of the rich data held in underlying battery, inverter and other equipment control systems.

By abstracting the underlying hardware, edge computing solutions from suppliers such as IOTech allow the vendors to choose and easily integrate energy products from a variety of equipment suppliers. By partnering with companies such as IOTech, these vendors can maximize energy density and create the most appropriate overall hardware and software solution for each of their customers or customer sites.         

VMblog:  Are companies really implementing AI/ML at the edge or is it just hype?

Foster:  AI/ML technologies are most definitely being used in edge environments so it's not just hype.  For example, IOTech is involved in a number of projects where AI/ML inferencing engines deployed at the edge are being used to process video feeds for object classification and identification for Retail Point of Sales (PoS) applications, or simply to count people in a venue management system.

Some use cases require this level of sophistication, but it can also be overapplied. In most cases simple rules or scripting engines can provide a lot of value at the edge, saving operational costs, improving safety and even generating new revenue streams. For sure some use cases are suited to AI/ML at the edge, but where the complexity is not needed companies are learning to keep it simple. There is still a lot of benefit that can be derived by measuring a few edge data values and automatically actuating when things get out of range.

##