Edge Mesh, as a new model, proposes that all computing tasks and data be shared using a meshed network of edge devices and routers, offering many advantages, such as distributed processing, low latency, fault tolerance, greater scalability and enhanced security and privacy.
These benefits are paramount for critical applications that require real-time processing, increased reliability or mobility support.
As we know, the Internet of Things (IoT) is revolutionizing industry by connecting all of its assets to operating systems through applications. Most IoT applications can be grouped around four types of functions: sensing, communication, computing and actuation. A single device does not usually support all the capabilities, so until now most systems used end devices to sense the environment, while communication and network responsibilities were assumed by gateways and computing was usually performed in a centralized cloud-like server, which sent the information generated, during processing to selected devices that later acted as executors.
However, this centralized computing model is not efficient for compute-intensive and time-critical applications such as Energy and Grid Operators, Utility Sector, the Water Industry as well as companies with critical asset for instance
If we look into Industrial IoT, a study from Juniper Research has found that the global number of Industrial IoT connections will increase from 17.7 billion in 2020 to 36.8 billion in 2025. As the number of devices increases, so does the volume of data and the importance of generating useful information. Computing is an important part of Industrial IoT, as it leads to the generation of new knowledge, which is used to optimize industrial processes much more intelligently. A good example of the latter, are the new intelligent industrial processes that have evolved to the point where they can understand the environment and act accordingly.
This new scenario of connected objects has led to the emergence of new management systems for edge devices, also known as edge nodes, applications and data, such as Barbara's Edge Platform. edge nodesBarbara's Edge Platform, which enable real-time response and management of highly critical assets in highly distributed environments.
Decision making in these scenarios is done within the network by sharing data and computation between devices instead of sending all data to a server. This new distributed system is changing the way centralized computing is done, where edge devices were only used for collecting and sending data to a server for processing.
Now with Edge Computing Platforms like the one of Barbara's, edge nodes are used to enable distributed intelligence in the Industrial IoT. Not only Edge Mesh is the new paradigm for distributed intelligence, it also enables "self-healing" capability, so that if a node fails in communication it can reroute around it, allowing the network to continue operating, and thus, increasing reliability.
How can the Edge Mesh model of distributed and cooperative computing respond to cloud problems? The cloud computing model presents 4 main problems: latency issues, security, privacy and scalability.
Edge Mesh has two main objectives:
Latency: Edge devices are increasing their compute, storage and communication capabilities every day, which has led to a new model of opportunistic cooperation, where edge servers can make use of surrounding devices for processing tasks. This solves one of the big problems of the cloud: overloading. Often, devices are overloaded with multiple tasks while others are underutilized, leading not only to an uneven distribution of tasks, but also to increased power consumption and latency.
Security and Privacy: by controlling data from its source location and therefore deciding what and when to send to the cloud, cybersecurity risks due to theft or improper access to information are reduced. Nevertheless, being a network of distributed resources, connected in many cases to critical elements, its security design, protection and monitoring is essential and this is of much greater relevance when Edge Nodes can operate the connected equipment,
Scalability: at the Micro-Edge level, the number of devices deployed can be very large (thousands to tens of thousands). Therefore the installation, provisioning and maintenance of Edge Nodes can raise the hidden costs of deployment to the point of being uneconomical. Especially in the case of industrial installations, which have extremely long lifetimes, it is essential to have tools that facilitate this lifecycle management of Edge Nodes in a remote, centralized and scalable way .
Distributed Computing and Edge Computing are born to address the problems of latency, mobility, security and the bandwidth bottleneck of traditional cloud computing. However, both have their benefits and drawbacks, but can work in a complementary way with each other to satisfy the multiple requirements of today's industrial applications.
A new model is then imposed, that aims to decompose applications into microservices and use resources both at the edge and in the cloud to alternatively satisfy the requirements of different applications.
A fully distributed model involves a huge management effort. The diversity of IoT devices and applications makes it almost impossible for a single model to satisfy all application requirements. This is solved by the Edge Mesh model that involves the integration of different systems.
Data interoperability is key for enabling distributed intelligence as we face distributed systems with synchronization, consensus, cooperation, heterogeneity of devices and applications issues.
Furthermore, In the industrial world there are no fully pervasive communications protocols or common data structures. It is therefore important that an edge computing mesh network deployment is based on open technologies, ideally standard or widely used by industry, that allow effective integration and evolution of different parts with the deployed infrastructure. Monolithic, closed solutions with high integration costs should be avoided.
The benefits of Edge Mesh such as distributed processing, fault tolerance, low latency, have great impact on companies with critical assets that are dispersed and generate data at high frequencies. One example is the use of Mesh Network of Edge Computing for Electrical Transformer Substations scattered across a country.
When it comes to digitize an Electrical Substation, several questions emerge:
How can the interoperability of the different electricity service providers operating at each Substation be ensured?
How to coordinate adjacent Substations?
How can data and equipment cybersecurity be guaranteed in such a widespread and critical environment?
In this environment, Edge Computing allows the integration of equipment, sensors and actuators to compute data in real time in a distributed manner, and is a better alternative to Cloud Computing. By implementing Edge Computing, Transformation Centers can process data locally from different sources, and make autonomous decisions quickly, without the need to go through centralized systems such as SCADA or Cloud. This, in the medium and low voltage grid, implies an immense leap in the operation and maintenance possibilities, with a great impact on improving costs, response times, scalability, continuity and reliability in the new Smart Grid paradigm.
Barbara's Industrial Edge platform is designed to deploy Artificial Intelligence at the Edge in a cyber secure manner. It enables the execution of Artificial Intelligence in the field, so that stakeholders can abstract from connectivity and communication frameworks, cyber security and IT infrastructure and focus on the execution and management of their algorithms.
Barbara offers a number of unique alternative technologies applicable to the digitization, virtualization and communication of Transformation Centers:
- Allows plug and play installation of Edge Nodes in Transformer Substations. These Nodes are capable of real-time processing of data coming from multi-manufacturer equipment, from the center itself or from adjacent centers ("Edge-Mesh
- It sends the processed results, or derived alarms, to the operator's systems via protocols close to operator systems via protocols close to the industry, such as OPC-UA, or new protocols like MQTT or API REST/HTTP.
- It also integrates other data from industrial systems such as SCADA, Historians, relational databases, for an automatic learning structure.
- Allows the execution of complex Artificial Intelligence algorithms trained outside the Node. These algorithms can be from different authors, since Barbara's technology isolates them in independent containers within the Node, in a microservices architecture.
This communication model breaks away from the traditional communications architecture, where Transformer Substations normally communicate via PLC or point to point fiber networks with the Transformer Substations immediately upstream or downstream, to move to a meshed communication architecture for the execution of local algorithms and tasks that can have a real-time response.
If you found the article of interest, and you want to knwo more, do contact us and request a personalised demo or consultation.