Wind River Carves out Its Position As an Independent Firm

What’s New with the IoT Supplier after Parting from Intel?

I recently caught up with Jim Douglas, president and CEO of Wind River, to discuss how the company’s been getting on since its recent split from Intel, and how it sees industrial Internet of things (IoT) going into 2019.

Under the tag line “When it matters, it runs on Wind River”, the company supplies operating systems, virtualization platforms, simulation technology and tools for industrial IoT systems. It specializes in non-stop, mission-critical operations often found in aerospace, defence, automotive, industrial, medical and telecom uses. And it recently underlined its credentials by becoming one of only a few suppliers whose products have been to Mars twice (see here).

Wind River was bought by Intel in June 2009 to help the latter build up its non-PC markets, such as embedded systems and mobile devices. Earlier in 2018 it was spun out and then acquired by private equity firm TPG Capital.

What’s Happened since Separating from Intel?

Wind River had been operating fairly independently within Intel, so the split was easier than some. A major outcome is that its new relationship with Intel is much more focussed, without the inevitable myriad of internal issues that exist in a large company.

Mr Douglas said he’s very pleased with the headway his company has made after being acquired by TPG Capital. The past few months have seen Wind River focus on setting itself up properly, and delivering against its initial objectives, which have progressed well. It’s now in a position to step up investment in its longer-term plans and, strangely, Mr Douglas also said the business may have readier access to funds than before, despite Intel’s deep pockets.

How’s the IoT Market Looking Heading into 2019?

The embedded computing market isn’t new. And data isn’t the new oil — it’s been generated within those systems for many years. What is new is the pace of connecting things, along with the ease of accessing, processing and analysing the data they produce.

But the most valuable new development is the use of embedded computing in digital transformation. In IoT, this will increasingly rely on pushing intelligence to the edge of a system.

With those trends as context, Wind River has three main top-level focus areas from here. The first is to make the economics of digital transformation work in industrial IoT. This relies on virtualization, so that multiple IoT “apps” or “workloads” can be run on the same machine, bringing flexibility and economies of scope to different types of machine. The second aim is to embrace the fact that there’ll be an assortment of hardware and software in use, and to make that easier to work with.

To meet these two goals, Wind River is putting a lot of effort into its Helix Virtualization Platform, a single software virtualization environment. The platform is in general availability and already adopted by important clients, but won’t be fully launched until early 2019. It forms the orchestration layer of Wind River’s “fluid compute” concept, with which it can improve efficiency and resilience by managing software workloads across devices automatically.

The third focus area for Wind River is the shift of the main source of differentiation in operations technology markets from hardware to software. Suppliers of operations technology need to become skilled in modern software and cloud practices, and this is a difficult and scary idea for many. It will require translating techniques from enterprise IT and cloud into the operations technology world.

This is an enormous task considering the fragmentation and complexity in embedded computing markets. Interestingly, Mr Douglas said Wind River is being helped by a rapid change in the mind-set of many of its critical infrastructure customers. Having been traditionally very cautious, some are now actively pushing Wind River for continuous integration and continuous deployment of new software updates. This is creating a level of demand-pull on the suppliers between Wind River and end customers.

How Does Artificial Intelligence Play into This?

Wind River sees an advance from automated devices to autonomous systems over the coming years, which will spur demand for significantly more computing power at the network edge. The company won’t compete by developing algorithms and taking part as an artificial intelligence specialist. Instead, Wind River aims to make it easier for developers to deploy their algorithms and neural models onto a wide range of hardware and software.

With other efforts like Intel’s OpenVINO and Amazon Web Services’ SageMaker Neo already having an impact on the market, Wind River recognizes that it’s a little behind, and is now making a strong push into building compatibility with a range of artificial intelligence frameworks.

Mr Douglas highlighted a shift here, which is set to become important in many industrial IoT and artificial intelligence systems: if a customer is embracing the continuous integration and continuous roll-out of software updates, as well as artificial intelligence models, how can it go through the full certification process that some industries require with each update?

For traditionally cautious industries, there are two main considerations when moving to this approach with software and machine learning. One is a greater reliance on testing the updates on virtual representations of the systems, known as digital twins. Another is that Wind River is already seeing a new willingness to deliver software updates rapidly, provided that it’s possible to roll back to a previous version very quickly and automatically if problems are found.

Wind River is in a really interesting position, carving out its place somewhere between the major cloud providers with their huge investments as they increasingly set the agenda in industrial IoT, and the traditional “box makers”, who have historically been the lead suppliers in those markets and are now having to get used to cloud, cloud computing at the edge and fog computing architectures.