Onteon for edge computing in manufacturing

This article is an overview on how to use Onteon for edge computing in manufacturing.

It references to “Edge Computing in the Context of Open Manufacturing” document, created by Open Manufacturing Platform – IoT Connectivity Working Group.

Reference use case in the OMP document distinguish 3 levels – cloud, edge and production asset level. In this article I will focus on edge level application as this is more specific than cloud level.

To understand the use of Onteon in this scenario let’s start with short introduction to Onteon.

Onteon is a platform/software stack to deploy and manage applications and native microservices. The difference to other solutions on the market is that Onteon works with containerized and non-containerized workloads the same way. It can operate also on OS level which brings additional functionalities and benefits.

To shed light what Onteon does, at this point I will just name what it includes: application provisioning, orchestration, runtime, load balancers, services communication, monitoring and log management.

At the infrastructure level on edge node there would be Onteon Node Manager (ONM) used as a guest system, running containerized and non-containerized workloads. It could be deployed on hypervisor or directly on operating system of a node. Onteon Node Manager offers providers for containers (Docker), JVM applications with runtime, and generic provider which manages all other polyglot applications and microservices.

OMP document describes three integration components: cloud connector, message broker/API gateway and production asset connectors. All three can be implemented as microservices managed by ONM communicating with each other. Of course all computation components can be implemented as containers or native microservices as well.

The integrated part of Onteon is service communication. Every Onteon Node Manager has internal and external load balancer built-in. This way Onteon delivers loosely coupled architecture for application to application communication, no matter if it’s inside the node or among the nodes. It allows to build logic and separation of data moving among production asset, edge and cloud level. Any north and south application level communication allow/deny scenario can be implemented.

All nodes communicate with each other and share the knowledge about applications and microservices deployed. Therefore, from communication perspective each node can serve as failover node for the other, providing appropriate applications and microservices are deployed there. Onteon will enable this automatically. To ensure services continuity production assets would need to have failover information to be able to try another route to production asset connector if the main node they communicate with will fail.

As reliability and high-availability is crucial for industry applications, Onteon have some mechanisms to achieve this goal. Apart from above mentioned failover of services to other nodes, Onteon Node Manager delivers in-service, non-disruptive node upgrades. Upgrading applications on the fly is not an issue in most architectures. However upgrading the node itself very often results in stopping the services on this node for the upgrade/reboot time. Onteon Node Manager can be upgraded on the edge device without stopping the applications and services running on it.

Also Onteon Node Manager can work autonomously without being connected with the “brain” part - Onteon Control Center. Onteon Node Manager has the knowledge which and how many applications should be working on it. It allows to work in offline situations, where communication between Node Manager and Control Center is broken. Although it is Onteon Control Center that takes care of nodes operability, Onteon Node Manager itself will try to reconnect to Onteon Control Center for maximum reliability.

Bandwidth and performance

Microservices and applications deployed and managed by Onteon can use any protocol. Onteon itself supports HTTP/HTTPS and binary protocol in the near future. So called adaptive communication allows registered services to talk to each other with the fastest protocol available. This is auto negotiation based on the available protocols common on both sides of communication. Any 3rd party can write its own protocol which can be handled by the applications working in Onteon environment.

When using other services with their protocols it is transparent for Onteon. For example MQTT subscriber using spring boot can be natively deployed and managed by Onteon, where no overload is added to communication by Onteon itself.

Most tools can operate only with containers. Container to container communication has its cost. This can impact response time and also number of requests served in a given period.

As Onteon allows to deploy non-containerized services and manage them the same way as containers, it can boost performance. Or from another angle we can say it allows to achieve the same with lower hardware resources of industrial PC or any edge gateway.

Security

All components of Onteon communicate with each other by HTTPS. Onteon Control Center, Onteon Node Manager, Monitoring and CLI. All custom applications and services can communicate with HTTPS. It can even be forced by Onteon to use HTTPS by applications to avoid non-encrypted traffic. Onteon has built-in Key Management Service with periodic key rotation mechanism.

Monitoring and log management

Onteon comes with the log and monitoring management in a package. It can be however connected to already used tools. As Onteon works on OS level it can be used for gathering events and logs not only from application, but also OS level.

Summary

According to IoT Analytics, platform hype is moving from cloud to edge. It is among top 10 trends in IoT for 2022. And the projected CAGR for edge computing market during 2022-2028 is 31.1% according to Valuates 2022 report.

Onteon can manage in a simple and efficient way different workloads: containerized, native binaries as well as script files. It works at the edge but also can be used on some other constrained devices.

According to IoT & Edge Developer Survey Report 2021 by Eclipse Foundation, three main concerns were security (46%), connectivity (38%) and deployment (31%). All three points are addressed in the latest release of Onteon. It provides many capabilities for edge computing out of the box.