[adamjcook]

I found this article pretty interesting (and detailed) in terms of VW's #NEWAUTO software strategy: Vw. OS (2025): Three Architectures, Android Automotive | AUTO MOTOR UND SPORT (auto-motor-und-sport.de)

Of interest, perhaps, is this passage for us existing ID series owners:

A word about the time-consuming and expensive triple work (1.1, 1.2 and 2.0): Updates between the systems (and thus also the vehicle platforms) are excluded. In other words, if you sit in a VW ID.3, you have no chance of the "real" VW. OS and version 2.0 in the car. But, and this is important to the Cariad developers: Many of the new features and improvements that are being developed for version 2.0, for example, will also be available in the 1.1 or 1.2 cars with a little delay.

Emphasis mine.

That is encouraging.

It also sounds like VW is forking Android Automotive for the E3 1.2 platform (and roundtripping Android Play Store access via a VW account).

Sincerely,

Adam J. Cook

 

[SunWizard]

This is more confirmation of the existing press reports covered in this thread:
Android automotive OS takes over the market, can we benefit?
Good news and it helps answer this a little: we will benefit "with a little delay". It will be interesting to see if we get the VW play store on our ID.4. This seems likely since that is where the new subscription features and apps will be available.

 

[ChristophW]

Interesting article, and quite a bit more spelled out than what we surmised and guessed up to now.
Pros: as you mentioned, backports of new functionality.
Not so great: ICAS1 and ICAS3 chips are not all that powerful and impose limitations. Bad choice on VW's part to go with those Siemens parts.
Bad news (pure conjecture on my part, but likely a business reality): When 2.0 is rolled out in 2025, expect far fewer and less frequent backports to our 1.1 version that is now old hat, boring, and subpar.

 

[adamjcook]

Pros: as you mentioned, backports of new functionality.

Bad news (pure conjecture on my part, but likely a business reality): When 2.0 is rolled out in 2025, expect far fewer and less frequent backports to our 1.1 version that is now old hat, boring, and subpar.

I sort of expected this, but it is nice to have some sort of confirmation. The existing E3 1.1 platform seems to basically now be the "tip of the spear" where VW can, theoretically, perfect their internal OTA process and new iterative development strategy, as well as, concurrently develop certain in-house software modules that will carry over to E3 1.2.

I would speculatively expect that at least some "moderate" measure of backports continue as long as necessary to maintain discernable residual or resale values in E3 1.1 vehicles.

Not so great: ICAS1 and ICAS3 chips are not all that powerful and impose limitations. Bad choice on VW's part to go with those Siemens parts.

I believe that the In-Car Application Server (ICAS) is a Continental product.

Anyways, I really think, more or less, VW was "forced" to go with ICAS as it corresponding closer to their existing product lifecycle familiarities in order to get these vehicles to market in a realistic timeframe even though VW, internally, might have strived for something a bit more enduring and more in-house right off the bat.

Herbert Diess was already aggressively (or at least not shy about) talking up vw.OS and "E3 2.0-like things" months before even the ID.3 launched so that was always my impression, for whatever that is worth.

Sincerely,

Adam J. Cook

 

[ChristophW]

You're right about ICAS being a Continental product, my bad. Same argument though, and actually worse since Continental is not natively a silicon hardware company as far as I know. I wonder if anyone knows what the chips are based on - ARM, MIPS, something else?

I agree with your assessment of how it might have gone down. Makes a lot of sense.

 

[adamjcook]

You're right about ICAS being a Continental product, my bad. Same argument though, and actually worse since Continental is not natively a silicon hardware company as far as I know. I wonder if anyone knows what the chips are based on - ARM, MIPS, something else?

I agree with your assessment of how it might have gone down. Makes a lot of sense.

I never got the impression that the Continental ICAS was inadequate from a compute standpoint, only that VW was already looking beyond it to something “more specific” to where they saw what is now coined E3 2.0 and this “Level 4 ready” business (the latter being as questionable to me as Tesla’s current “Full Self Driving” nonsense).

Although there is obviously some considerable flexibility in the definition, Continental always described the ICAS as a “high performance computer”.

There was a very detailed Continental slide deck floating around about the nitty-gritty details and architecture of ICAS that I saw about a year and half ago. I will see if I can find it.

Theoretically and speculatively, VW should be able to improve E3 1.1 quite considerably on ICAS. We shall see on that.

Sincerely,

Adam J. Cook

 

[Nai3t]

If the old way of thinking is "you got the car that you bought," meaning not to expect too much in the way of upgrades, but now VW is telling us "expect upgrades," then the sane expectation would be to accept the general capabilities of the car we bought, but maybe a bit more spit 'n polish as we walk the update path together. No amount of updates will convert the ID.4 into an autonomous mobility robot, nor increase the screen resolution, nor add two additional window switches, but I'm perfectly content (and expect!) to see what refinements they can work out with the hardware we have.

 

[Salish.4]

You're right about ICAS being a Continental product, my bad. Same argument though, and actually worse since Continental is not natively a silicon hardware company as far as I know. I wonder if anyone knows what the chips are based on - ARM, MIPS, something else?

I agree with your assessment of how it might have gone down. Makes a lot of sense.

It might help this discussion to understand that it is the German company, Continental, that has and will for some time provided both the onboard hardware (servers) and software for the VW ID series. See Start of the Electric Era: New VW ID.3 E-model Equipped with Continental Technologies
VW claims 2 lead servers for the ID (see my comments above) while Continental claims that they are providing one HPC ("high-performance computer"). I am not sure how to reconcile this.

 

[Nai3t]

This is where I find it gets interesting.

Continental has a subsidiary called Elektrobit which focuses on software development.

Elektrobit has a joint venture with Cariad called e.solutions that worked on the ID.3 software (and presumably our ID.4 software since it's all one-in-the-same).

Cariad is, of course, VW's subsidiary software team.

Putting all of this together, my impression is that this resource-sharing structure allows VW to bring Continental's software expertise into the VW camp, and allows a progressive transition towards VW's all in-house software development team to press ahead with its 2025 target without any hard cut-over transitions.

 

[Salish.4]

This is where I find it gets interesting.

Continental has a subsidiary called Elektrobit which focuses on software development.

Elektrobit has a joint venture with Cariad called e.solutions that worked on the ID.3 software (and presumably our ID.4 software since it's all one-in-the-same).

Cariad is, of course, VW's subsidiary software team.

Putting all of this together, my impression is that this resource-sharing structure allows VW to bring Continental's software expertise into the VW camp, and allows a progressive transition towards VW's all in-house software development team to press ahead with its 2025 target without any hard cut-over transitions.

Thanks for the information. I have been trying to put the Continental side together with the VW side. How this transition fares is critical to VW. What we have seen is that the technology side (most of this car) is having a challenge.

 

[Salish.4]

Thanks for the information. I have been trying to put the Continental side together with the VW side. How this transition fares is critical to VW. What we have seen is that the technology side (most of this car) is having a challenge.

Wow! Just visited Elektrobit's website, specifically Trends - Elektrobit
For me, this expanded view of what is behind these new vehicles is an eye opener and I wonder how I (and this forum) is going to move the conversation past trailer hitches into understanding and influencing what these machines are becoming. It is good to see Electrobit dedicate a section to security as these cars and our phones mesh.

 

[Spin]

Elektrobit has a joint venture with Cariad called e.solutions that worked on the ID.3 software (and presumably our ID.4 software since it's all one-in-the-same).

Seems to me they could have done a better job.

 

[Nai3t]

Well we all know the ID.3 software was a disaster for VW in that it caused the car to be delayed and was incomplete at launch. Many of the press releases from these various companies are dated mid 2019 and even 2020. If you'll recall, early 2020 saw "sky is falling" news reports stating ID.3 software development was rushed and poorly planned, various systems inside the car weren't playing together well, and the car was due to launch that summer.

It sounds like either VW completely underestimated this task, or they hastily got this team together late in the game to catch up. Or both!

 

[Mike Bouck]

My day job is writing software and this doesn't bother me - the hardware is always getting better. What matters to me is does the current ID.4, as a car, make financial sense to me right now and the answer is yes. In 4-5 years from now the tech will be better in every measurable way - that is a given. Does that mean I'll lose 95% of the value of my current ID.4? I highly doubt it - there will always be a market for well-built, reliable, well-treated used cars.