There are three types of computing platforms within the edge to cloud continuum: Functional edge, edge computing, and cloud computing.
The functional edge is best described as a place where people and things collaborate to reflect their combined intelligence and processes in the making of a product.
People means individuals in different business disciplines who are involved in production. For example, in a manufacturing facility:
- Operators set up and run
- Plant managers oversee the overall plant performance
- Equipment engineers handle procurement, deployment, and life cycle management of assets machines
- Technicians focus on maintenance and repair
- Quality engineers ensure quality
Things represent physical assets on a factory or shop floor that are used in making the product. They come in different forms, including:
- Manual assets that are managed by people, such as trolleys or carts
- Semiautomated assets that are run by a combination of control systems and people, such as a machine that processes raw material but still requires people to load and unload material or for operation
- Automated assets that are managed by a control system, such as automated conveyors, automated mobile robots, and IIoT devices
Edge computing is described as bringing information processing closer to the functional edge, where things and people produce or consume that information. It includes two types of computing: Far edge and Near edge.
The far edge is a segment of edge computing that uses architecture models to carry out communication, data aggregation, and near real-time computations at the end-point device locations.
Far edge includes three types of deployment models:
- Device edge: Focuses on acting upon the data that is generated by the functional edge and creation of user experience. As an example, Programmable Logic Controllers (PLCs) are used for the collection of telemetry data and control of semiautomated and automated assets of the functional edge. People use end devices, such as smart pads and human-machine interfaces, to interact with the assets.
- Compute edge: Sets up a compute platform closer to the device edge. It is used for processing and storing data, creating user experiences, communicating, and collaborating with the device edge and gateway edge to manage the functional edge.
- Gateway edge: Bridges the device edge and functional edge with the compute edge through network connectivity. It stores and forwards data that is collected from the device edge to the compute edge and back and can also send command and control data from the compute edge to the device and functional edge.
Near edge computing supports diverse architecture models to onboard new servers, applications, and network segmentations to carry out large amounts of data computation, storage, and communication locally, but farther from end-point devices.
Near edge includes two deployment models:
- On-premises data center, which provides complete ownership and control of edge computing and facility network infrastructure management and is typically chosen for:
- Data security
- Accessibility to physical hardware
- Managing the sprawl of physical servers
- Tailoring for performance to meet the needs of far edge and functional edge
- Agility to scale based on need
- Regional data center, which supports distributed computing capabilities and network infrastructure management for two or more facilities that are based in the same geographical location. It is typically chosen for:
- Avoiding the sprawl of servers across facilities
- Reducing cost of ownership for individual factories
- Using co-location and hyperscaler cloud provider offerings, with servers and operations that are designed to rapidly deliver scalable cloud services, process large amounts of data, and integrate with other systems
Cloud computing provides access to scalable distributed computing and network management capabilities through cloud-based data centers. These centers are typically managed and governed by hyperscaler cloud providers.
Cloud computing provides the advantage of flexibility and efficiency to meet the needs of an edge ecosystem through As-a-Service models. These models are typically Infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS) used for development and deployment of use cases across multiple facilities with agility.
Like edge computing, As-a-Service models come with different deployment models such as public and hybrid clouds for meeting the needs of the edge ecosystem in terms of reliability and scalability. However, they come with concerns around response latency, bandwidth, and trustworthiness in processing and storing data that is sensitive to industrial companies.
The main types of cloud computing are:
- Public cloud, which delivers the advantage of cloud computing through hyperscaler-provided computing platforms for industrial organizations. It is primarily used for cost efficiencies in managing the workloads of individual facilities, and the availability of software solutions from well-established vendors.
- Hybrid cloud, which delivers a single optimized computing infrastructure to run the workload of an edge ecosystem. The hybrid cloud is realized by connecting public cloud to on-premises or regional data centers to meet the unique requirements of individual facilities.