Operational technology (OT) networks have been around for decades, successfully managing industrial and manufacturing systems that monitor and control complex physical processes. These include everything from robots on a production line to the valves and pumps in oil and gas pipelines to the traffic management systems for trains. During this time, the underlying technology in OT networks has evolved from TDM to ATM to IP/MPLS and Segment Routing.
Ensuring operational continuity is critical in OT networks because when an outage occurs, the potential consequences can be catastrophic. Manufacturing grinds to a halt, oil and gas don’t get distributed, and trains can’t depart or arrive at their destination. Not surprisingly, these networks require low latency and the highest levels of reliability. They also need to be highly secure. In many cases, OT networks are completely isolated from the internet to ensure they are protected from cyber threats.
With the advent of private wireless, enterprises are now connecting more intelligent equipment and sensors using the Internet of Things (IoT). Controlling these devices and collecting data from them are business-critical capabilities, which means that reliability and deterministic performance are foundational requirements, especially as these networks expand. But OT networks are more than just collection networks; they also continue into the data center.
Today, as businesses across every industry press ahead with digital transformation initiatives, information technology (IT) networks are evolving along the same lines as OT networks, adopting many of the same fundamental characteristics.
Unlike the previous generation of applications, which typically ran on bare metal servers, new OT applications are typically “cloud native.” The cloud native approach has many advantages, including disaggregating the application from the server infrastructure, giving operators more deployment options, and enabling them to scale applications on demand.
These applications are decomposed into containerized microservices and Kubernetes clusters resulting in traffic flows both east-west within the data center, then north-south to the network and to devices and users. Cloud native and Dev-ops are two sides of the same coin, providing more flexibility and iterative deployment of new functionality. This contrasts with the previous generation, where applications were more stable, but also largely static.
Also at the application layer, there is emergence of Artificial Intelligence (AI), machine learning and deep learning which will use an enormous amount data collected from the machinery, controllers, sensors, or “things” delivered over the OT field network to the AI clusters deployed in data centers centrally, in the cloud or at the edge of the network.
To support cloud-native OT applications, the data center fabric needs to meet stringent requirements for availability, quality of service (QoS), security and resiliency. Achieving the OT service levels means having resiliency baked into the network design via Active-Active multihoming, link-aggregation and multipath networking. Having the assurance that the required bandwidth is available, and the right priority is assigned to the application traffic, is also critical. For example, if network resources are shared with other bursty traffic, then it’s crucial to ensure that high-priority operational traffic flows are not affected.
Data center fabrics are also adopting the technology of the WAN such as Ethernet Virtual Private Networks (EVPN). They are using Multi-Protocol – Border Gateway Protocol (MP-BGP) as the control plane in the network and EVPN Active-Active multi-homing redundancy to enhance network resiliency. EVPN with MPLS tunnels in the WAN is extended to the applications in the data center with EVPN over virtual extensible local-area network (VXLAN) tunnels.
OT networks have always been mission-critical which means that network platforms are held to a higher standard of hardening, reliability, and resiliency. Now these networks are being extended to within the Data Center and combined with more typical IT traffic. As enterprises adopt new technologies, IT and OT convergence will progress as the attributes of OT and IT worlds continue to influence each other. One essential requirement for these convergent OT and IT networks is security but warrants a separate blog … to be continued.