Migrating Test and Simulation Workloads to the Cloud

Overcoming legacy dependencies

Test and simulation technology is experiencing breakaway innovation thanks to the benefits of cloud networking and the tone set by market leaders. Some benefits of cloud networks for test and simulation include the ability to address limitations of on-prem environments, with regard to burst-capacity, resource queuing, and compute power capabilities. In conjunction, large players have sent the market a clear signal by directing investment to update legacy software for cloud compatibility and to develop new cloud-native solutions. cloudSwXtch from swXtch.io is uniquely positioned to support the test and simulation industry’s cloud journey. cloudSwXtch is a virtual switch that offers the ability to add many high-performance features to cloud networks that support migration as well as performance optimization.

Let's dive into how cloudSwXtch supports engineers in the test and simulation market. In this article, we'll cover:

• Cloud-based test and simulation in the industrial world.

• Multicast and broadcast dependencies and their impact on migrating test and simulation workloads to cloud.

• Value of a “lift and shift” migration approach to preserve legacy systems.

• cloudSwXtch features that support cloud-based test and simulation.

Cloud-based test and simulation in the industrial world

In industrial engineering, test and simulation workflows contribute to a range of design and operations activities. These workflows strive to quantify performance under different scenarios that represent real-world conditions and to predict and optimize outcomes before changes are rolled out to a live production environment. The benefits of cloud network tools and capabilities can be seen across several fronts:  

• Consistency – Engineers can spin-up VMs using scripts to deliver dependable and consistent environments. These VMs can be created to represent the performance parameters of an asset or environment and be deployed repeatably for each required simulation or project. Environments can even be pre-configured to utilize the same IP structure so that each additional VM in a new VNET is easily deployed and each network is identical to every other.

• Project turnaround – Multiple simulation VMs can be setup in parallel to take advantage of the underlying elasticity of cloud resources. This reduces the time it takes to run a set of simulations and ultimately supports faster project turnaround. Engineers can even run simulations in faster than real-time if desired.

• IT investment – Engineers can effectively navigate cloudbursts for peak IT demand and utilize specialty integrated-circuit boards to match compute requirements. Gone are the days where companies need to invest in additional hardware or software licenses to accommodate peak demand scenarios.

What this looks like

Automation system simulation and virtual controls testing

Test and simulation software is essential when designing and commissioning automation systems and controls. Systems engineers and control room operators rely on these test and simulation environments to beat-up control equipment logic in different operational scenarios before systems are deployed, assets are modified, or new assets added.

In the case of power plant controls setup, simulations mimic data from thousands of distributed sensors and power-meter nodes, representing dynamics of control in the face of various fluid, electrical, and other materials flowing through the plant.

Using the same industrial protocols, data from the simulation appears on the HMI (human machine interface) and allows the operator to turn virtual knobs and change values to see how the simulated plant would respond. After designs have passed test and simulation, they get integrated into the controls system logic and deployed for operators to use.  

This process is much faster and easier to scale than in the past when engineers relied on manual calculations or slower systems to test and deploy their designs. Thanks to cloud based virtual machine (VM) simulations, engineers can now analyze and test their designs in real-time or faster.

Multicast and broadcast dependencies and impact on cloud migrations

Many on-prem test and simulation platforms, especially related to automation, were built using multicast or broadcast-dependent protocols. These include Ethernet/IP, BACNet, EGD, and other UDP-based protocols. These protocols serve many functions, are pervasive in a large installed equipment base, and are unfortunately blocked from cloud networks. Benefits of these protocols include:

• Supporting real time endpoint cooperation across large distributed systems and utilizing multicast to reduce point-to-point communication.  Especially as organizations attempt to mimic on-prem environments in the cloud, virtual machines need to be coordinated in much the same way. Each simulator can range from a few to hundreds of VMs and when data needs to be shared simultaneously with many or all VMs, multicast and broadcast protocols are required.

• Discovering device and resources that are part of a multicast group is more efficient than sending separate packets to each device. Multicast allows engineers to identify devices on the network and complete bulk delivery of data.

• Reducing network capacity requirements by condensing the amount of data that must be transmitted in many copies via unicast.

Why companies prefer lift and shift for cloud migration

When organizations plan to migrate existing applications to the cloud, they have a few options, including re-coding or re-architecting, but many prefer lift and shift.

• Supporting long-standing assets: Maybe it was built in the 1980s and is still performing! Legacy applications work a certain way on-prem and companies may not want or be able to make changes to the simulation application: There are proprietary models and protocols, extensive customization, and specific methods of data storage and retrieval that are not easily to replicate.

• Recertifying and revalidating applications requires a considerable effort and cost: Moving simulation applications to the cloud involves significant work and expense.

• Internal teams and customers have a level of built-up familiarity with existing platforms, especially in industries where an installed base lasts for decades.

As one client recently told us, “We are trying to live with the fact we can't make any changes and we want to figure out how we can build this inside of Azure.”

cloudSwXtch features that support cloud-based test and simulation

• Multicast and broadcast protocols: Distribute multicast and broadcast traffic from on-premises to the cloud and between different cloud networks.

• Fanout: Accelerate workloads and reduce overhead by sending packet flows to all end-points in a group using multicast. cloudSwXtch can convert unicast and broadcast packet flows to multicast and distribute it to all desired end-points in a group.

• Reliable transport: Address loss characteristics associated with link quality and certain protocols with several available features, including SRT, 2022-7, high-availability, and Forward Error Correction (FEC).

• Kubernetes: Containerized network support that allows multicast and other protocols to traverse pods and containerized applications.

• High Performance: Users can deliver millions of packets per second to an unlimited number of endpoints with sub-millisecond latency.

• Existing Infrastructure: Add cloudSwXtch to existing infrastructure to extend applications into cloud networks to achieve the same look, feel, and performance in the cloud as on-prem.

‍