Auto DevOps

  • Comments posted to this topic are about the item Auto DevOps

  • Hi Steve, this is a quick thought on Auto software.  It is becoming an amazing high tech industry.  I recently heard a discussion from a gentlemen who has the left ear of the US Government's Safety Traffic administration about how fast it is starting to evolve. Basically the thinking is the IOT is actually going to spearheaded by a lot of the automobile companies as they refine vehicle to vehicle and vehicle to infrastructure communications. We already have Cadillacs in Detroit that are warning their drivers of upcoming traffic light changes in operation today by GM's global technical center. These vehicles are working towards an environment where they will be creating a self replicating, self healing autonomous networking environment. For instance, if a car slips on a curve on a highway, it will be able to communicate that there is an icy patch on the highway so that the vehicles behind it will be able to adjust their traction control systems. But if a car erroneously relays that message on a hot summer day, then the networking environment will be smart enough to isolate that vehicle from the network and display a message to that vehicle's dash that a sensor is off and probably needs servicing.

    However, I think  your concern that the systems will all be integrated with the entertainment systems and powertrain modules is something that is already being considered.  Infotainiment systems will be over the air networked for real time streaming with 5GL networks (still being developed) that are being anticipated to be faster than today's fastest cable systems for their internet connections.  And in the meantime, the vehicle to vehicle to vehicle and vehicle to infrastructure communications will be very well defined with small real time messages with such small amounts of data (we're talking just a few bits almost) that some of the security concerns will already be addressed in terms of just hardware isolation alone.

    The features that control the vehicle will be so completely different from infotainment systems that there is very little current danger of any coupling or sharing of any data due to completely different architectures and processing capabilities. Even simple hazard switches have micro-code in them these days and a modern vehicle now has more source lines of code than a F16 fighter jet.

    But from a simple  view point of what I am seeing, your concern should not be an issue (although I do confess I'm more involved with traditional IT, and do have not access to future product plans).

  • If there's one thing we've learned in a few decades of writing software, it's that large monolithic systems are a mess.

    FTFY

  • I totally get the potential value of IoT enabled assembly plants, commercial transport vehicles, or shipping containers. But when it comes to embedding IoT in consumer facing products (like automobiles or refrigerators), it seems that most of the discussions revolve around the challenges involved or perhaps the potential use all that data might have for the manufacturers, either directly or by selling it to 3rd parties. I mean, what benefit all this has for the consumers themselves is less obvious.

    If IoT can improve the efficiency of the manufacturing or fulfillment process, resulting in more affordable products, then great. But high-end IoT enabled devices that cost twice as much and contain "features" nobody asked for in the first place; I have doubts about that business model.

    Maybe I'm just playing Devils Advocate, or maybe I'm just getting old and cranky, but the IoT evangelists really need to find a better way to sell IoT to the broader public, not just investors or an audience at TED Talk, if they want their business ventures to turn a profit and succeed in the long term.

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • I have zero interest in IoT being added to my car.
    It is simply another point of failure.  Expensive to troubleshoot and fix.
    My wife's 2015 GM products 'infotainment' system is a hot mess, is confusing to use and has already proven to be problem prone. 
    Automakers are known for ignoring manufacturing issues/defects on their vehicles.  What makes anyone believe that these same manufactures will spend the time and money developing patches for older software on their cars?  You could/will end up with serious security issues in your car and with no recourse.  
    Maybe auto manufactures will allow you to buy a software subscription so that you can get updates to your car.  No thanks!

    I like my older car without all the integrated 'things'.

  • Just wait until government regulators start issuing automobile recalls based on IoT equipment failure or security issues. Some may see it as an unprofitable risk. Shareholders and the corporate board of directors may sour on IoT and start complaining: "Mr. CEO, please remember we're an automobile company, not a software and data company."

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • podmate - Monday, July 31, 2017 9:17 AM

    I have zero interest in IoT being added to my car.
    It is simply another point of failure.  Expensive to troubleshoot and fix.
    My wife's 2015 GM products 'infotainment' system is a hot mess, is confusing to use and has already proven to be problem prone. 
    Automakers are known for ignoring manufacturing issues/defects on their vehicles.  What makes anyone believe that these same manufactures will spend the time and money developing patches for older software on their cars?  You could/will end up with serious security issues in your car and with no recourse.  
    Maybe auto manufactures will allow you to buy a software subscription so that you can get updates to your car.  No thanks!

    I like my older car without all the integrated 'things'.

    And you can't upgrade it as you see fit. Many newer cars can't replace the entertainment, because it's embedded/linked with other systems.

  • kjg48359 - Monday, July 31, 2017 6:31 AM

    Hi Steve, this is a quick thought on Auto software.  It is becoming an amazing high tech industry.  I recently heard a discussion from a gentlemen who has the left ear of the US Government's Safety Traffic administration about how fast it is starting to evolve. Basically the thinking is the IOT is actually going to spearheaded by a lot of the automobile companies as they refine vehicle to vehicle and vehicle to infrastructure communications. We already have Cadillacs in Detroit that are warning their drivers of upcoming traffic light changes in operation today by GM's global technical center. These vehicles are working towards an environment where they will be creating a self replicating, self healing autonomous networking environment. For instance, if a car slips on a curve on a highway, it will be able to communicate that there is an icy patch on the highway so that the vehicles behind it will be able to adjust their traction control systems. But if a car erroneously relays that message on a hot summer day, then the networking environment will be smart enough to isolate that vehicle from the network and display a message to that vehicle's dash that a sensor is off and probably needs servicing.

    However, I think  your concern that the systems will all be integrated with the entertainment systems and powertrain modules is something that is already being considered.  Infotainiment systems will be over the air networked for real time streaming with 5GL networks (still being developed) that are being anticipated to be faster than today's fastest cable systems for their internet connections.  And in the meantime, the vehicle to vehicle to vehicle and vehicle to infrastructure communications will be very well defined with small real time messages with such small amounts of data (we're talking just a few bits almost) that some of the security concerns will already be addressed in terms of just hardware isolation alone.

    The features that control the vehicle will be so completely different from infotainment systems that there is very little current danger of any coupling or sharing of any data due to completely different architectures and processing capabilities. Even simple hazard switches have micro-code in them these days and a modern vehicle now has more source lines of code than a F16 fighter jet.

    But from a simple  view point of what I am seeing, your concern should not be an issue (although I do confess I'm more involved with traditional IT, and do have not access to future product plans).

    Perhaps, but anytime we share data, or receive data, there are the possibilities for interference or hacking. I worry that the extreme example you show (ice in summer) isn't the issue. It's someone hacking and showing no ice in winter, or showing lots of ice and causing other issues.

    It's  unexpected hacking that can occur, because of the known ways we have cars respond. Maybe not a big deal, but a concern. Right now there is a central computer that controls many aspects of the system. If they share a network, who's to say that someone can't get an entertainment system to send a note to the transmission to put it in park? Or an alert along the network. I think you're minimizing the vandalism of people and their incredibly creativity.

  • 20% of BMW's employees are software engineers? WOW, that's amazing! That's getting pretty near the point where BMW isn't so much as an automobile company, but a software company that makes automobiles.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • We escaped a Ford Escape after it spent more time in the shop for software issues than we drove it. Nothing like having system failures at 70 MPH to realize that cars are over computerized...

  • Rod at work - Monday, July 31, 2017 11:07 AM

    20% of BMW's employees are software engineers? WOW, that's amazing! That's getting pretty near the point where BMW isn't so much as an automobile company, but a software company that makes automobiles.

    My guess is that they're working on their own version of self-driving technology.

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • Already there have been issues with manufacturers trying to patch mechanical things through software because it's much cheaper.

    One example I am familiar with because of knowing some drivers, is a patch that was applied to some models of the Jeep Grand Cherokee because the electronically actuated transfer case (the system that manages the 4wd) was not always fully engaging and the vehicle could roll in park. The patch was to modify the settings in software to make sure the system was engaged instead of changing out components ... however this caused all manner of problems with some systems that had been fully operational but perhaps slightly out of spec (can't get a straight answer from the factory). Hence people's 4W Low range would periodically disable and go into limp mode, requiring a visit to the dealer. Dealers would replace circuit boards... sometimes it would work, sometimes the same problem would occur later with no predictable pattern. Some people had gone back multiple times.

    There has developed a 'black market' in circuit boards from junkyards which have never been patched (once it's patched, it appears impossible for users to reset it and the dealers can't or won't).

    My wife's older model GC is purely mechanical. Sure it takes a little more effort to engage the gears, but at 20 years of age and well over 200K miles, there are none of those problems.

    The problem in the industry as a whole is that all sorts of things that should be left mechanical have been converted into a complex array of sensors and actuators (who on earth needs electronic door latches? What happens if you get caught in a flood?). Failures of a circuit board are difficult to diagnose and astonishingly expensive to replace (repair is not an option). Cars built around touch screens become a real problem when the screens fail, and replacement is ridiculous. I am not so much worried about hacking, but expense of repair is a serious concern with aging.

    ...

    -- FORTRAN manual for Xerox Computers --

  • jay-h - Tuesday, August 1, 2017 1:54 PM

    The problem in the industry as a whole is that all sorts of things that should be left mechanical have been converted into a complex array of sensors and actuators (who on earth needs electronic door latches? What happens if you get caught in a flood?). Failures of a circuit board are difficult to diagnose and astonishingly expensive to replace (repair is not an option). Cars built around touch screens become a real problem when the screens fail, and replacement is ridiculous. I am not so much worried about hacking, but expense of repair is a serious concern with aging.

    All good examples of issues, but not the full story for me. I don't think that things are mechanical or electronic solely. Why not electric door latches? It's convenient and helpful. This could be a safety feature as well (lock doors as your drive, or to stop someone from getting in. However, they can't be just electronic. I should be able to open these from inside for safety reasons. I can't believe this isn't the case in every car.

    Touch screens are an issue for the driver. I think they're fine, but they should be on the passenger side, or in the rear. Certainly my wife should be able to enter navigation while we're driving. It's sometimes useful. However, me, as a driver, shouldn't get distracted and I should have tactile responses. One reason I avoided a Lexus is that many features use a mouse interface. I shouldn't be moving a mouse to some spot while driving. I like a mechanical interface that I Can turn and click, and not spend more than a couple seconds making a choice, just like I take a couple seconds to check a mirror.
    Pilots have  also reported issues  as too many items move to  touch screens, a quick radio change or heading alteration can be distracting if it's not a mechanical switch.

    I think it's fine that  software  can be updated, but it must be  secure, and limited. And not  locked to  specific methods. The same way that I shouldn't have features locked to the manufacturer. I should be able to upgrade portions of the car, if appropriate.

  • Steve Jones - SSC Editor - Tuesday, August 1, 2017 2:44 PM

    All good examples of issues, but not the full story for me. I don't think that things are mechanical or electronic solely. Why not electric door latches? 

    I perhaps should have been more clear on that. For years we've had cars with door latches that could be locked etc electrically, that's not so much a problem (though even that wasn't very nice when a Buick locked me out in the cold while I was scraping ice off the windows) but full electronic where on some vehicles there is NO mechanical connection (and that worries me if something like getting stuck in a flooded road shorts the CAN bus). There have been a number of recalls recently on some popular cars where the latching becomes erratic or fails to latch properly.

    On a whim recently I wandered into a Tesla showroom and the friendly salesman urged me to get inside. "Just touch the recessed door handle". Really do we need touch sensitive door handles?

    I agree with you that touch screens are a real potential safety issue (don't know why the regulators seem to ignore this), tactile physical switches can be operated without taking eyes off the road.

    ...

    -- FORTRAN manual for Xerox Computers --

  • My x5 has all-electric latches. I had one die and the passenger couldn't get out when we parked.

    Less than pleasant, and this has to be a safety issue.

Viewing 15 posts - 1 through 14 (of 14 total)

You must be logged in to reply to this topic. Login to reply