Why we need to evolve from ‘software-defined’ to ‘self-driving’?

In the digital economy, user requirements and workloads are simply exploding—in real time. While core IT management needs have not changed, they have been made more complex as cloud and mobile apps become pervasive. To keep workflows optimal, great user experience is key.

Nowadays, a string of public and private cloud platforms serve as the main driver for ubiquitous data storage, access and processing. Where latency is an issue, edge computing takes care of real-time data collection and all the processing needed to make instant use of the data before channeling it back to the Cloud for subsequent warehousing.

Together, Cloud and Edge computing have facilitated a better user experience when implemented via best practices. What are the core IT management elements required to provide an even better user experience and business agility? How do we transform the edge with autonomy? Where do we go from here?

The journey to an autonomous edge These questions were addressed in a recent DigiconAsia virtual fireside chat with edge computing solutions firm Stratus Technologies’ VP of Strategy and Product Management, Jason Andersen. Andersen explained how businesses can make use edge devices to streamline and rationalize IT technologies, making them resilient, simple to manage remotely and with non-tech-savvy staff, and most importantly, granted the autonomy to use AI and machine learning to perform routine tasks with minimal human intervention.

In these times of lockdowns, recessions and business struggles, how can enterprises quickly use edge computing to streamline their IT technologies for survival and growth? Does making edge solutions autonomous (self-driving) add value to the digitalization effort?

In order to quickly use edge computing to enhance multiple IT technologies, Ansersen said, there are three areas to focus on:

1. Cloud connectivity
Cloud solutions for managing data are great, but you need specialized, highly trained personnel to manage and service the technology at the edge. The complexity and cost of orchestrating the availability of maintenance personnel and equipment can introduce risk and reduce response times.

By using autonomous edge systems, the large IT footprint can be reduced. This can be achieved via intelligent control software that reduces the need for skilled manpower or complex computing equipment.

 Additionally, using edge solutions such as virtualization and containerized code can make connectivity to the cloud solutions seamless, thereby creating a more agile overall Cloud-to-Edge infrastructure.

When properly implemented, autonomous edge computing reduces the cost of using the Cloud, optimizes the size and usability of the edge, and results in a well-balanced, cost-efficient infrastructure.

2. Hybrid OT at the edge

Operational technology can be made more agile via edge solutions for virtualization and standardization of control across multiple OT facilities. This can also reduce costs and increase security via device cloning and advanced systems backup capabilities.

For example, remote monitoring capabilities reduce the need for on-site human supervision, allowing personnel to be deployed more productively. Being autonomous and easy-to-use, edge systems can be operated even by non-technical staff, thereby boosting resilience and response time.

Finally, while failure redundancy is often in place in OT infrastructures, challenges can be encountered when a failed system needs replacement. With autonomous edge solutions, the control system would alert personnel with predictive maintenance alerts and can even order the right replacement parts or schedule the most qualified personnel to service any malfunction.

3. Smarter autonomous manufacturing

Robotics and production lines often include custom-made equipment requiring specialised logic controllers, mechatronic assemblies (skids) and so on. These are costly to acquire and maintain, and require skilled personnel to manage.

Autonomous edge computing promotes the use of generic components that are controlled by software logic instead. With such a control system, remote monitoring and management can be handled from a single edge compute platform. This reduces the need for skilled manpower, lowering acquisition, maintenance and replacement costs. It improves worker productivity while the programmability of software control can be leveraged to enhance the versatility of robotics and assembly lines.

Andersen believes that edge computing has come into its own with AI and ML, and businesses that embrace self-driving edge computing will be able to use data efficiently to improve the user experience at all levels of commerce.

Industries such as operational technology, manufacturing, aerospace, semiconductors, automotive, Food & Beverage and public transportation are primed for the use of autonomous edge. Citing an example, Andersen said: “Instead of alerting personnel that something is wrong, the autonomous edge controller would have also initiated the order of spare parts and scheduled the appropriate technician to attend to the problem.” Such predictive maintenance is a trademark of the autonomous edge.

When planning for the transition to a self-driving edge platform, the “next steps” should always involve the human touch. “Who’s driving innovation in the company? What are the current skills sets of the teams? What are the future needs of the business? IT may be providing the voice of innovation, but they may not have enough control to do it.”

This is why IT teams need to work closely with the rest of the organization and understand how to best optimize processes, before they can digitalize processes with the right autonomy and efficiency.

Edge computing vs autonomous edge
Answering questions from the fireside chat moderator and the internet audience about the difference between autonomous and standard edge systems, Andersen noted that edge systems can be made autonomous via the following characteristics:

1. Rationalization: The simplification of IT technologies and processes with simple-to-use edge devices that employ AI and machine learning to minimize manual control.

2. Standardization: the elimination of customized components and manual onsite control in any edge system, in favor of remotely controlled, generic components managed via virtualized control logic and made simple to operate and maintain onsite by laypeople. This enhances system resiliency and response time to mitigate operational anomalies.

3. Analytics and autonomy: Edge devices that can offload data processing from the Cloud in situations which require real-time or instantaneous analytics. Also, such an edge system has smart logic that streamlines the management of numerous IT technologies into a single control platform that can be handled by remote personnel anywhere in the world via Cloud connectivity.

Andersen reiterated that an autonomous edge solution requires close and thorough communication between operations people and IT teams. “Only when the former group is able to convey their multifaceted needs and requirements to the IT group, can edge systems be designed for optimal outcomes,” he concluded.

Optional box story

Autonomous edge computing at play
In a world ravaged by the COVID-19 pandemic and propelled into widespread e-commerce and digital transformation, even real-time, resilient edge computing has to move in sync with business-critical mandates to survive amidst budget constraints and haste.

By empowering the edge with autonomous and self-driving capabilities, businesses can tap into the cloud-edge continuum with greater ease and improve user experience. How?
In Jan this year, IDC research indicated that edge computing costs and complexities can increase with scale, and this problem can be managed with the implementation of AI and ML. This level of ‘self-driving’ edge computing ultimately increases user experience at every level. For example:

In transportation: The most well-known example of the concept of self-driven edge computing has to be found in autonomous vehicles. AI and ML require super low latency edge devices to provide real-time data for predictive decisions in moving the vehicle and its occupants safely and comfortably to the destination. Conversely, sensors and computers alone will be insufficient to achieve truly safe, secure and autonomous vehicular behavior without elements of AI and ML.

In hospitality: Voice recognition, smart lighting and climate control; entertainment options and customer service—all these aspects can be greatly enhanced with smart real-time sensors and machine learning to cater to the whims of guests in any hospitality situation.

In manufacturing: According to IDC, as manufacturing undergoes digital transformation, physical assets can be emulated with a “digital twin” made up of virtual data that must be stored, processed, and analyzed. The latency has to be consistently small and reliable, and edge infrastructure holds the key to making the digital twin emulations self-driving.

In retail: Customer experience can be monitored by AI via an edge computing back end, transforming how the retail experience with intuitive sales information, personalization features and loyalty perks.

In Operational Technology (OT): As IT merges with legacy OT infrastructure, the digitalized OT system presents a larger attack surface to criminals. This is where AI and ML can be implemented at the edge to enhance vigilance, surveillance and preemptive analytics for better physical and security.

In human resources management: It is already well-known that the future workplace will maximize the creative and intellectual contributions of workers while minimizing mundane duties using AI and ML. Shortages of skilled manpower, challenges in retaining and grooming talent, will be mitigate with heavy dependence on AI and ML being implemented in the workplace via satisfaction-monitoring sensors and audits. Human error will be kept in check, while workplace morale and job satisfaction will form the crux of excellent worker performance across the board.