Perfecting the technological aspects of connected car systems may prove the easy bit for vehicle manufacturers; it’s working out a profitable and practical business model for mass production that’s the real challenge.
Connectivity is presenting today’s OEMs with a double whammy in that investment requirements are rising while profits are declining. Most vehicle manufacturers are investing tremendous amounts of money – billions of dollars in some cases – in the software space. At the same time, basic vehicle profit margins are under attack. With the incoming regulations around safety, security and fuel economy, some forecasts call for a 1-2% drop in average profit per vehicle over the next five to ten years. The key to surviving the squeeze could lie in software strategy.
Some OEMs today spend between US$50m and US$80m for a single instance of an infotainment system. They will generally develop separate systems for their upmarket brands or models, their mid-market offering and the lower-end offering. What if they could take the same basic framework, harden it, test it and reuse it across all of their various price classes, but just add enough differentiation on top to appeal to the consumer? The potential savings in time and cost would be enormous.
Plugging the software gap
“Today we are moving into a software-related industry, and the vehicle manufacturers are entering a space that they don’t necessarily understand,” observed Marques McCammon, General Manager, Connected Vehicle Solutions at Wind River.
At the end of the day, OEMs are basically aiming to accomplish four targets, which McCammon breaks down into the acronym ACRU: Abstract their dependencies at multiple levels in the way that they handle software; Consolidate their compute resources so they don’t have to simply keep adding more controllers and more engineers in order to execute effectively; Reuse the software that they’ve already generated and invested in; and Update the software and the experience to the consumer to provide long-term value.
The belief that OEMs need to take more control of the software and know from whence it’s coming is valid, but considerable time and attention has to go into looking at how they execute that part of the business
Some incumbents have been trying to develop all the necessary expertise in-house but find themselves up against Amazon, Google and Apple, the likes of which live in the software domain full time and move through it much more dynamically. “The notion of just setting up software teams or going out and buying software is probably not going to prove to be sufficient in the long run,” McCammon told Megatrends. “The belief that OEMs need to take more control of the software and know from whence it’s coming is valid, but considerable time and attention has to go into looking at how they execute that part of the business.” These companies need to invest only where it will yield the most profit and value to shareholders while still remaining dynamically competitive in the marketplace.
Focus on the framework
Some brands have been trying to address the software stack at all levels, meaning they are developing everything from the very root framework level all the way up to the applications that the consumer sees. This requires significant investment, some of which could be avoided without any impact on payback. One solution could be to create a level of separation in the approach to software.
Wind River suggests separating the software framework in the middle from its dependence on the computing side, including things like the microprocessors and the silicon on which the software runs, as well as the applications at the top layer that the consumer sees. The focus is squarely on the framework in the middle. “We try to find ways to standardise the framework across the industry to ensure a predictable performance. This also allows us to control costs,” said McCammon. “We encourage our customers to let us make that framework extremely strong so they can put their trust in it. They can then focus investment at the top level where the consumer sees the value.”
While this may represent a shift in thinking for automotive players when it comes to software, it’s not a million miles away from how they already approach car production. The framework of a car, or the chassis, is generally reusable at some level. “OEMs spend considerable time on developing a basic platform to give it reliability and durability. The more frequently they can use it, the more money they save,” he pointed out. “We can do the same thing with software, but it requires a different thought process.”
Wind River has already been working with Honda on such an approach with its Android-based infotainment system. The aim is to gain scale and speed.
Redistributing responsibility
If collaborations among OEMs, Tier 1 suppliers and software specialists start early enough, it is possible to redistribute the roles and responsibilities involved in the software design development. It is unlikely to be practical for every OEM to staff thousands of software engineers, but they could more easily staff several hundred and restrict their focus to those customer-facing areas. They would then look to the software industry to provide the frameworks and the Tier 1 suppliers to assist with integrating these frameworks.
“It represents a disconnect from the current supply chain model, where the Tier 1 provides all the software locked into a box and then hands it over to the OEM,” conceded McCammon. “In that model, the OEM does not have sufficient control of its destiny. The alternative is that the OEM does all the work in-house, but then it ends up spending money on aspects that don’t return value. The brands then lose the scale and the shared business risk model. We are trying to help the industry turn traditional thinking on its side.”
OEMs spend considerable time on developing a basic platform to give it reliability and durability. The more frequently they can use it, the more money they save
ACRU broken down
These changes become all the more pressing as companies shift from pilot volumes to mass market. Moving to production is a completely different discipline from operating in the bubble of a small scale trial. “When a company starts to move towards production, the change has to shift from sizzle to steak. Everyone needs to start looking at how to achieve scale and speed in the execution along with reliability and confidence in the delivery,” he noted. “They also need to protect themselves against potential for risk.”
This is where ACRU comes into play, starting with that abstraction away from dependencies on hardware. “If I’m not dependent on one particular ECU or one particular microprocessor in order to make my system work, then I can benefit from supply chain competition and drive down costs. This should also help me achieve speed,” McCammon commented.
Consolidation kicks in quickly as well. In the experimentation phase of autonomous vehicles, for instance, most OEMs run cars with super computers in the trunk, cooled by huge fans because they run extremely hot. That’s simply not possible in a production environment.
The savings from the ‘reuse’ aspect are easily apparent. In the early stages of experimentation, software models for automated driving features are very likely to be built on PCs. Once these are validated, engineers go back and redesign them in more production-oriented software. That whole approach is very wasteful. “We need to provide the industry with frameworks in which they can start developing and experimenting from day one,” said McCammon. “We need to make it so that as they evolve, they’re reusing the same software base over and over. There is less that needs to be re-architected and rebuilt.”
The final aspect, updating, is critical due to the evolving nature of software. “Because software continues to evolve, we have to make sure that we have the ability to touch it continuously. When we talk about updating vehicle software, we’re only thinking about harvesting value post deployment, but the reality is we can use that update philosophy from the time that we start designing until the time that we start the next design. From cradle to cradle,” said McCammon.
A living, breathing thing
One of the biggest obstacles to realising this general approach is a misunderstanding about the nature of software design. It cannot be delivered like other hard components in the vehicle. “When we hold software to the same delivery pattern, the same production cycle and production schedule, resulting in a mismatch,” he warned. Hardware matures in a very different way to software. This goes back to the idea that software is constantly moving and changing.
The software specialists realise how organic software is. “It is like a living, breathing thing,” emphasised McCammon. Today everyone is trying to extract new value from it or make it stronger. In automotive, that’s not always possible due to the hard and fast production schedules. With the shift towards mobility as a service, that changes. “That has a direct impact on the economics and the quality of the automotive industry,” he added. “To realise that, we have to pivot our thinking as an industry to intercept the way that world functions and operates.”
This article appeared in the Q2 2018 issue of Automotive Megatrends Magazine. Follow this link to download the full issue
