Toggle Search
   Arm Enables

Workload Consolidation on Arm

For new factories, converging multiple single-workload machines into fewer ‘do-all’ devices has an extensive list of benefits.

Workload Consolidation

Workload consolidation, the concept of converging multiple single-workload machines within a factory into fewer ‘do-all’ devices, has an extensive list of benefits. It’s a concept being embraced by new, greenfield factories that don’t have (and don’t want to have) a floor full of disparate devices of varying age, function, space and power requirements to maintain.

Without such constraints, greenfield factories can focus on increasing productivity while reducing both physical and carbon footprints on their journey to autonomous operation.

The industrial sector represents a huge installed base of Arm-based chips. As OEMs strive to standardize their architecture from top to bottom, Arm continues to invest heavily in augmenting this scalability through workload consolidation.

Device manufacturers looking to consolidate multiple devices into a new multi-function device will find great benefit in retaining the same underlying Arm compute architecture using accelerated experience-based development and existing tool chains. This migration can also have a direct impact on the total cost of ownership, both in the hardware and software stack, not only for the OEM but also throughout the supply chain.

Arm offers a suite of IP and technologies that support and enable virtualization, real-time deterministic performance, functional safety and security.  With a vast ecosystem, the Arm partnership makes achieving workload consolidation a reality.

Substantial increase in compute performance

Naturally, when multiple applications or workloads are consolidated, the compute performance needs increase: performance is proportional to the level of integration. The breadth of the Arm portfolio allows the selection of ‘right sized’ compute right up to the highest performance requirements for these developments. In cases where a large number of functions are consolidated, these requirements can be significant—potentially even requiring server-class platforms such as those in the Arm Neoverse family.

One consideration in migrating low-power compute workloads on to higher performance platforms is thermal dissipation. The common way to cool electronic systems is through the use of fans, but this then creates a secondary issue: factory floors can be dirty and dusty, and these fans can get clogged up over time and fail. This causes the equipment to go into thermal shutdown, resulting in an expensive ‘line stop’ situation.

Arm’s portfolio of energy-efficient processors offers renowned performance-per-watt capabilities. This can vastly reduce the need for fans, enabling developers to rely more on passive cooling solutions leading to higher reliability, lower power consumption production lines for the factory operator and greatly reduced noise pollution on the factory floor.

Workload consolidation within edge gateways

Edge servers are increasingly found within factories. While their primary purpose is to gather, filter and then backhaul vast amounts of data from endpoint IoT sensors to the cloud, the powerful compute these devices offer can also be put to use in performing tasks previously given to other disparate devices.

For example, analytics that drive more localized operational optimization and decision making. This would happen prior to passing the data up the hierarchy for further analysis related to productivity, material management and integration. This could then feed into supply chain management including delivery updates and customer related notifications.

Again, these types of platform often consolidate multiple functions including protocol translation, security, communications and machine learning (ML) analytics to name but a few. This can be realized through different software technologies including both virtualization and containers offering maximum flexibility for any deployment.

Safety and security in workload consolidation

Security is paramount in covering a range of vulnerabilities and threat models in industrial systems. As Operational Technology (OT) and Information Technology (IT) networks begin to merge, there is the potential to expose a greater attack surface outside the factory. This places additional security requirements on each piece of equipment. A workload consolidation platform may need to implement specific security hardware features and could also run security applications within a separate virtual machine. Arm has many assets that enable true end-to-end security including technologies such as Trust Zone, CryptoCell and PSA Certified.

Security goes hand in hand with safety, which focuses on the physical environment in this context. For example, ensuring that a manufacturing robot is not put in a position in which its operation—whether functional or malfunctional—could cause damage to itself, other assets or to the humans working in the factory.

Arm safety technologies are key to solving the problems of deploying safety efficiently in industrial applications. The way in which safety capable products are developed is critical, and this is why Arm has a rigorous Arm safety process and range of Arm Safety Ready IP and relevant software that are developed for use across multiple applications and functional safety standards, including IEC61508.

In a consolidated system approach, different functions may have differing functional safety requirements.  This is often referred to as ‘mixed criticality’ and can add complexity to the overall systematic design of the product.

For more information on functional safety, read Functional Safety in Industry: Thinking Outside the Cage.

Addressing mixed criticality in workload consolidation

As workload consolidation is undertaken, an inherent challenge that needs to be addressed relates to their differing requirements.  Some workloads will be deterministic in nature such as a Programmable Logic Controls (PLC) application, used to monitor and control processes and systems in everything from washing machines to elevators to heavy machinery. The level of deterministic or real time requirements vary upon the application.

This is often referred to at the cycle time within an industrial application. In some applications of workload consolidation, running a real-time operating system (RTOS) on an application processor such as the Arm Cortex-A family offers an acceptable level of determinism—this is called ‘soft real-time’. Other applications may need tighter cycle times and more dedicated ‘hard real-time’ compute, offered by real-time processors such as the Arm Cortex-R family.

Others will be more generic, such as a Human Machine Interface (HMI), which handle the exchange of information between a compute system and a human operator.

There may also be applications that have a functional safety requirement. In any consolidated system, these individual critical requirements must be guaranteed. But through separation and freedom of interference provided by the hardware and software, these ‘mixed criticality’ applications can co-exist successfully on the same platform, operating both efficiently and safely.

The Arm ecosystem’s critical role in workload consolidation

To take full advantage of the benefits that workload consolidation offers, multiple ingredients must converge seamlessly. Operating systems, whether general purpose or real time, and hypervisors are key for successful implementations. This is where Arm’s strong and vibrant ecosystem plays a critical role in providing and supporting these elements – working together to make workload consolidation possible.

Workload consolidation brings multiple benefits throughout the industrial supply chain and this is exactly why the deployment of workload consolidation is accelerating.

With so many existing workloads running on Arm hardware already, workload consolidation is a logical step forward in performance, functional safety, security and real-time processing. With the support of our partners, Arm and its ecosystem can ease development burden and help move to this new, streamlined way of working.

Back to top