Building boundaryless cloud edge solutions

A blog by Myke Miller, managing director and dean of the Cloud Institute, Deloitte Consulting LLP; David Linthicum, chief cloud strategy officer, Deloitte Consulting LLP; Diana Kearns-Manolatos, senior manager, Deloitte Center for Integrated Research

In order to guide organizations on how to harness the full potential of modern-day computing, we spoke with edge computing pioneer Mahadev Satyanarayanan (Satya), the Carnegie Group University professor of computer science and Deloitte Cloud Institute fellow. For our exclusive interview and detailed perspective, watch for our Deloitte Insights article, “The Edge of cloud: A discussion with Mahadev Satyanarayanan, the Carnegie Group University Professor of Computer Science,” which publishes on March 25. Here, we’ll share his and our thoughts on the present opportunities and future challenges related to adopting cloud-edge architectures.

A four-tiered computing model

When thinking about developing a tiered computing architecture that brings together cloud, edge, mobile, and batteryless platform technologies, think through the different components of the system individually, and provide the flexibility in your cloud-edge architecture to adapt data and workload flows as requirements change.

In respect to cloud, Satya explained, “The clearest and simplest way that I have found to think about the different components of a modern computing system is to think of it as a tiered architecture in which different tiers are responsible for delivering key properties that we depend upon. The cloud is fundamentally the place in the system for precious or archival data that you want to be sure is around. It could be a private cloud, it could be a public cloud; that’s a separate decision, but the cloud basically is where you have the highest level of longevity of data; it is the most secure location. Think of it like a bank where you keep money or keep valuables, like a deposit vault. So that’s the cloud.” The cloud may be a preferable location to send continuously streaming data that doesn’t require low-latency computing (such as financial transactions data for end-of-day clearing and settlement), depending on a system’s computing latency requirements and data storage and analytics cost considerations.

In the case of mobile and edge technologies, Satya continued, “The edge is where you use the data, and in today’s terminology, the edge could be a desktop, but more likely, it’s going to be a smartphone or possibly a wearable device because mobility has been a very big driver of modern computing. The challenge is: I have this faraway cloud, and I have these mobile devices that impose certain demands on computing. These are things like size, battery life, etc. It is the price you pay for making things mobile, but then, instead of a bank vault, you get a Swiss Army knife (a screwdriver, a knife, and half a dozen other things). Where a single function may be suboptimal, if mobility is your goal, carrying the Swiss Army knife is the right way to do it.” Manufacturing and industrials employing smart factories and digital supply chains dependent on real-time sensor data and the constraints of batteries, and where cloud data centers may be too far away to power real-time computing needs, could benefit from proximity solutions (mobile and edge computing) with historical data to be offloaded to the cloud to manage cost and power broader enterprise AI initiatives.

When understood individually for their unique benefits and constraints, organizations can then advance to thinking through a tiered architecture strategy that brings together cloud, edge, and IoT technologies to support data analysis and AI programs. These Intelligent Edge Network strategies bring together mobile, edge, cloud, and batteryless platforms to power individual or stackable and interrelated use cases.

As boundaries blur across computing architecture tiers, firms will need to think through configuration management and orchestration approaches that are flexible across these tiers to allow for adaptability as functional requirements change with the evolution of hardware, software, and algorithms. As smart factories, utilities, and industrials and other industries explore edge-native applications and hybrid edge cloud strategies, being able to partition data and workloads across architecture tiers will become increasingly important. Additionally, being able to build an economic business case for a cloud-native versus edge-native application strategy will be a foundational success factor for organizations.

As Satya says, “The next chapter in computing is going to be both the creation of this edge computing infrastructure worldwide and its use to create brand-new edge-native applications, applications you just simply couldn’t create if you only have to rely on the cloud.”

To learn more about those applications and the advancements of new solutions being offered by providers, watch for the full article, “The Edge of cloud: A discussion with Mahadev Satyanarayanan, the Carnegie Group University Professor of Computer Science,” which publishes on March 25.