While we were at Embedded World last week, Intel announced two initiatives focused on the “connected car” concept:
Within today’s automobile, the channels for end vehicle differentiation are changing. Now when a prospective buyers steps into the vehicle, the first thing most people look at and play with is not the seat or the steering wheel but the in vehicle infotainment systems. As you would expect, this has quickly translated to increased R&D investment from the OEMs and infotainment systems are now emerging as platforms for personalization and driver customization through downloadable content.
This new venue for product extensions and value added services is clearly appealing to both OEMs, consumers, as well as other ecosystem constituents. However, there is obviously a significantly different additional dynamic within the automotive sector, as compared to say that for mobile phones, because of the substantial number of safety-critical subsystems for which the failure could lead to operator injury or even loss of life.
As such, automotive OEMs are now being forced to look for new ways to enable this connectivity, but doing so in a way that doesn’t jeopardize the security of the entire systems. Additionally, the growing amount of electronic functionality and increasing number and heterogeneity of subsystems in today's automobiles is already complicating the development process in the sector. Now many of these OEMs are actually reevaluating a technology that could help them address both of these issues – which is virtualization.
Not only does this provide a mechanism for consolidating the number of subsystems as a means to control BOM costs, but it can also afford OEMs the ability to securely run multiple OSs and subsystems over the same piece of silicon. As you can see from these findings from our recent research, more and more automotive engineers report that they are using or expect to use virtualization in their current solution.
What is especially interesting here is that, going back two or three years, I had a number of conversations with automotive manufacturers to support a research project that I was working on at the time. When the topic of virtualization came up, the OEMs were very quick to dismiss it as not appropriate for automotive designs. Clearly, their opinions have now changed, but this past experience leads me to another important question:
Given what is rapidly becoming a consumerization of the automotive industry, can the existing ecosystem with that automotive culture of conservatism support a consumer electronic-like velocity of change?
Or will the increasing level system connectivity provide a new venue for non-automotive OEMs (such as Apple or HTC) to swoop in and capture the next level of in vehicle experience differentiation?