Incident: GM's Delayed Response to Critical Car Hacking Vulnerability

Published Date: 2015-09-10

Postmortem Analysis
Timeline 1. The software failure incident involving the hack of a General Motors' 2009 Chevy Impala by security researchers at the University of California at San Diego and the University of Washington happened in 2010 [51221].
System 1. General Motors' 2009 Chevy Impala's OnStar dashboard computer system [51221] 2. Eighth generation OnStar units in GM vehicles [51221]
Responsible Organization 1. General Motors (GM) - The software failure incident was caused by a vulnerability in GM's OnStar dashboard computer software that allowed for remote exploitation, leading to potential control of the vehicle's functions [51221].
Impacted Organization 1. General Motors (GM) [51221] 2. National Highway Traffic and Safety Administration [51221]
Software Causes 1. Lack of security preparation in the entire industry of Internet-connected cars, leading to vulnerabilities in GM vehicles [51221] 2. Inability of automakers five years ago to fix hackable bugs in their vehicles' software promptly, unlike companies in the desktop and server world [51221] 3. Delayed response and lack of readiness by GM in 2010 to address the threat of car hackers, resulting in a prolonged period of vulnerability [51221]
Non-software Causes 1. Lack of security preparation in the entire industry of Internet-connected cars, leading to a delay in addressing vulnerabilities in vehicles' software [51221]
Impacts 1. Chrysler issued a software fix before the research was made public, and the National Highway Traffic and Safety Administration launched an investigation, leading to a 1.4 million vehicle recall [51221]. 2. General Motors took nearly five years to fully protect its vehicles from the hacking technique disclosed by researchers, leaving millions of GM cars and trucks vulnerable to the attack [51221]. 3. GM's delayed response to the software vulnerability in its vehicles highlighted a lack of security preparation in the entire industry of Internet-connected cars [51221]. 4. GM eventually developed the ability to push "over-the-air" updates to its vehicles to patch vulnerabilities, but the process was not without challenges [51221]. 5. The researchers' successful demonstration of the hack on GM vehicles raised concerns about the industry's readiness to address cybersecurity threats in connected cars [51221].
Preventions 1. Implementing a robust incident response and update system for Internet-connected cars within the automotive industry could have prevented the software failure incident [51221]. 2. Enhancing security preparation and capabilities within automakers to address and fix hackable bugs in vehicles' software promptly, similar to how tech companies like Microsoft and Google handle software vulnerabilities [51221]. 3. Developing the ability to push over-the-air updates to vehicles for software patches and fixes, as demonstrated by GM in response to the OnStar hack [51221]. 4. Emphasizing cybersecurity practices within the auto industry to address emerging threats and vulnerabilities promptly, as seen in GM's improved response to cybersecurity concerns in recent years [51221].
Fixes 1. Implementing over-the-air software updates to patch vulnerabilities in vehicles' software [51221] 2. Enhancing cybersecurity practices within the auto industry to address threats of car hacking promptly [51221] 3. Developing a universal incident response and update system for Internet-connected cars [51221]
References 1. Security researchers at the University of California at San Diego and the University of Washington [51221] 2. General Motors (GM) [51221] 3. National Highway Traffic and Safety Administration [51221] 4. GM chief product cybersecurity officer Jeff Massimilla [51221]

Software Taxonomy of Faults

Category Option Rationale
Recurring one_organization, multiple_organization (a) The software failure incident related to hacking a car over the Internet and exploiting vulnerabilities in the OnStar system happened again at General Motors (GM). A group of security researchers at the University of California at San Diego and the University of Washington had privately disclosed the exploit code to GM in 2010, targeting GM's 2009 Chevy Impala. GM took nearly five years to fully protect its vehicles from the hacking technique, indicating a failure to address the vulnerability promptly [51221]. (b) The articles mention that the incident involving hacking cars through vulnerabilities in the OnStar system highlighted a lack of security preparation in the entire industry of Internet-connected cars. The researchers pointed out that automakers, including GM, were not equipped to fix hackable bugs in their vehicles' software promptly, similar to how tech companies like Microsoft and Google address software vulnerabilities quickly. This lack of readiness in the automotive industry to address cybersecurity threats was a broader issue affecting multiple organizations within the industry [51221].
Phase (Design/Operation) design, operation (a) The software failure incident related to the design phase can be seen in the case of the car hacking technique targeting GM's 2009 Chevy Impala. The security researchers discovered a vulnerability in the OnStar dashboard computer that allowed them to remotely exploit the system, gaining control over various functions of the vehicle, including tracking, engaging brakes at high speed, and disabling brakes altogether. Despite the researchers privately disclosing the exploit to GM and the National Highway Traffic Safety Administration in the spring of 2010, GM took nearly five years to fully protect its vehicles from the hacking technique [51221]. (b) The software failure incident related to the operation phase is evident in GM's delayed response and patching process to protect its vehicles from the OnStar exploit. Despite efforts between 2010 and late 2014 to shield its vehicles from the attack method, GM's security measures were not fully effective until they silently pushed out a software update over its Verizon network to millions of vehicles with the vulnerable Generation 8 OnStar computer starting in November 2014. The delayed timing of the patch and the challenges in updating the software on the OnStar computers indicate operational difficulties in addressing the security threat promptly [51221].
Boundary (Internal/External) within_system (a) within_system: The software failure incident discussed in the articles is primarily attributed to factors originating from within the system. The failure was related to a vulnerability in GM vehicles' software that allowed for remote exploitation, enabling hackers to take control of various vehicle functions, including brakes and transmission [51221]. The delay in addressing this vulnerability was not due to external factors but rather to the lack of security preparation and capabilities within the automotive industry at the time. GM's slow response to fully protect its vehicles was a result of the organization not being optimally structured to address cybersecurity concerns in 2010 [51221]. (b) outside_system: There is no explicit mention in the articles of the software failure incident being caused by contributing factors originating from outside the system. The focus is primarily on the internal factors within the automotive industry, such as the lack of security preparation and capabilities, that led to the vulnerability in GM vehicles [51221].
Nature (Human/Non-human) non-human_actions, human_actions (a) The software failure incident occurring due to non-human actions: The software failure incident in the article was primarily due to vulnerabilities in the software of GM vehicles that allowed for remote exploitation by hackers. The attack targeted the OnStar dashboard computer and could control various functions of the vehicle, such as tracking vehicles, engaging brakes at high speed, and disabling brakes altogether. GM took nearly five years to fully protect its vehicles from this attack, highlighting a lack of security preparation in the industry of Internet-connected cars [51221]. (b) The software failure incident occurring due to human actions: The delay in fully protecting GM vehicles from the hacking technique was attributed to a lack of security preparation in the industry of Internet-connected cars. The researchers who discovered the vulnerability privately disclosed it to GM and the National Highway Traffic Safety Administration in 2010, but GM took nearly five years to patch the software vulnerability. GM's chief product cybersecurity officer acknowledged that the organization was not optimally structured to address the cybersecurity concern at that time [51221].
Dimension (Hardware/Software) software (a) The software failure incident discussed in the articles primarily relates to software vulnerabilities rather than hardware issues. The incident involved a hack on GM vehicles through their OnStar dashboard computer, which was exploited by security researchers [51221]. The vulnerability allowed attackers to remotely control various functions of the vehicle, such as tracking, engaging brakes, and disabling brakes. The delay in patching the software and the challenges faced by GM in addressing the security flaw highlight the software-related nature of the failure incident. (b) The failure incident is attributed to software vulnerabilities in the OnStar dashboard computer of GM vehicles. The security researchers were able to exploit these vulnerabilities to gain remote control over the vehicles, demonstrating a flaw in the software system [51221]. The delayed response by GM in addressing the software vulnerability and the subsequent patching process underscore the significance of software-related factors in the failure incident.
Objective (Malicious/Non-malicious) malicious (a) The software failure incident described in the articles is malicious in nature. Security researchers were able to hack into a General Motors' 2009 Chevy Impala through a remote exploit targeting its OnStar dashboard computer. The attack allowed the researchers to have complete control of the car, except for the steering, enabling them to track vehicles, engage brakes at high speed, and even disable brakes altogether. The researchers privately disclosed the exploit to GM and the National Highway Traffic Safety Administration in 2010, but GM took nearly five years to fully protect its vehicles from the hacking technique [51221]. The incident involved a chain of hacker attacks integrated into a single exploit, starting with a phone call to the Impala's OnStar computer triggering a vulnerability known as a "buffer overflow." This initial attack allowed the hackers to take control of the OnStar computer's data connection and penetrate the car's CAN bus, which controls various functions like brakes and transmission [51221]. The delayed response from GM in addressing the vulnerability in its vehicles, the challenges faced in patching the software, and the need for over-the-air updates to fix the issue highlight the malicious nature of the software failure incident [51221].
Intent (Poor/Accidental Decisions) poor_decisions (a) The intent of the software failure incident related to poor_decisions: - The software failure incident involving the hack of GM vehicles was partly due to poor decisions made by the automakers and the industry as a whole in terms of security preparation for Internet-connected cars [51221]. - GM's delayed response in fully protecting its vehicles from the hacking technique was not due to negligence but rather a lack of security preparation in the industry of Internet-connected cars [51221]. - The researchers argue that GM's failure to fully protect its vehicles from the attack did not reflect negligence but rather a lack of capabilities in the industry to fix hackable bugs in vehicles' software [51221]. - GM's slow response to the vulnerability in its vehicles was attributed to the fact that the organization was not optimally structured to address cybersecurity concerns in 2010 [51221]. (b) The intent of the software failure incident related to accidental_decisions: - The software failure incident involving the hack of GM vehicles was not due to accidental decisions but rather a lack of security preparation and capabilities in the industry of Internet-connected cars [51221]. - The delayed timing of the patch by GM was not accidental but rather a result of the organization not being ready in 2010 to deal with the threat of car hackers [51221]. - The failure to fully protect GM vehicles from the attack was not accidental but rather a reflection of the industry's lack of readiness to address cybersecurity concerns in the past [51221].
Capability (Incompetence/Accidental) development_incompetence (a) The software failure incident related to development incompetence is evident in the case of General Motors' delayed response to the car hacking technique discovered by security researchers from the University of California at San Diego and the University of Washington. Despite being privately disclosed to GM and the National Highway Traffic Safety Administration in the spring of 2010, GM took nearly five years to fully protect its vehicles from the hacking technique targeting its OnStar dashboard computer [51221]. (b) The accidental software failure incident is demonstrated by the flaw in GM's attempt to block the attack on its cellular network in 2011. Even after putting in place a measure to block data connections from OnStar computers to unauthorized servers, a flaw existed in the fix that allowed the exploit to work again in certain situations. The researchers found that one in every 10 or 12 times they restarted their Impala, its OnStar registered with the Verizon network in a way that failed to prevent it from connecting to a malicious server, unintentionally leaving the vehicles vulnerable to the attack [51221].
Duration temporary The software failure incident related to the vulnerability in GM vehicles to hacking was temporary in nature. The vulnerability persisted for nearly five years from the time it was privately disclosed to GM and the National Highway Traffic Safety Administration in the spring of 2010 until GM was able to push out a software update over its Verizon network to millions of vehicles with the vulnerable Generation 8 OnStar computer between November 2014 and the first months of 2015 [51221]. This indicates that the software failure was not permanent but rather existed for a specific duration due to certain circumstances.
Behaviour crash, omission, timing, value, other (a) crash: The software failure incident described in the articles can be related to a crash. The incident involved a vulnerability in GM vehicles' software that allowed hackers to remotely exploit the OnStar dashboard computer, gaining control over various functions of the vehicle, including the brakes and transmission. This loss of control over critical vehicle systems due to the software vulnerability can be considered a form of a crash in the system [51221]. (b) omission: The software failure incident can also be linked to omission. Despite efforts by GM to patch the vulnerabilities in later versions of OnStar, the security measures implemented did not fully prevent the exploit in vehicles using the vulnerable eighth-generation OnStar units. This omission in fully addressing the vulnerability left potentially millions of vehicles on the road susceptible to the attack [51221]. (c) timing: The timing of the software failure incident is also significant. GM took nearly five years to fully protect its vehicles from the hacking technique after it was privately disclosed to the company and the National Highway Traffic Safety Administration in the spring of 2010. The delayed response in addressing the vulnerability highlights a timing issue in the software security measures taken by GM [51221]. (d) value: The software failure incident can be associated with a failure in value. The vulnerability in the OnStar system allowed hackers to exploit the software, gaining control over critical vehicle functions such as tracking vehicles, engaging brakes at high speed, and disabling brakes altogether. This incorrect performance of the software, allowing unauthorized control over vehicle functions, represents a failure in the intended value of the software [51221]. (e) byzantine: The software failure incident does not exhibit characteristics of a byzantine failure, which involves inconsistent responses and interactions within a system. The incident primarily revolves around a specific vulnerability in the OnStar system that allowed for remote exploitation of vehicle functions, rather than erratic or inconsistent behavior within the system [51221]. (f) other: The software failure incident can be categorized under the "other" behavior as well. This incident highlights a failure in the industry's security preparation for Internet-connected cars, showcasing a lack of readiness to address software vulnerabilities in vehicles. The incident also underscores the evolving nature of cybersecurity practices in the automotive industry and the need for improved incident response and update systems to address software vulnerabilities effectively [51221].

IoT System Layer

Layer Option Rationale
Perception embedded_software The software failure incident discussed in the articles is related to the embedded software layer of the cyber physical system that failed. The failure was due to contributing factors introduced by embedded software error. The incident involved a vulnerability in GM vehicles' OnStar dashboard computer software that allowed hackers to remotely exploit the system, gaining control over various functions of the vehicle, including tracking, engaging brakes, and disabling brakes [51221]. The researchers were able to exploit a vulnerability in the OnStar computer's software through a series of hacker attacks integrated into a single exploit, demonstrating the critical role of embedded software in the cyber physical system failure [51221].
Communication connectivity_level The software failure incident described in the articles is related to the communication layer of the cyber-physical system that failed. The failure was due to contributing factors introduced by the network or transport layer. The incident involved a vulnerability in GM vehicles' OnStar dashboard computer that allowed hackers to remotely exploit the system, gaining control over various functions of the vehicle, including tracking, engaging brakes at high speed, and disabling brakes altogether [51221]. The attack method involved exploiting a vulnerability in the OnStar computer's data connection, which allowed the attackers to penetrate the car's CAN bus, the network of computers controlling various vehicle functions [51221]. Additionally, GM attempted to address the vulnerability by implementing network protection measures, such as blocking data connections from OnStar computers to unauthorized servers. However, flaws in these network protection measures allowed the exploit to still work, indicating a failure at the network or transport layer [51221]. Therefore, the software failure incident in this case was related to the connectivity level, specifically due to contributing factors introduced by the network or transport layer.
Application TRUE The software failure incident described in the articles was related to the application layer of the cyber physical system. The failure was due to contributing factors introduced by bugs and vulnerabilities in the software that allowed hackers to remotely exploit the OnStar dashboard computer in GM vehicles, leading to potential control over various vehicle functions such as brakes and transmission [51221]. The exploit involved a chain of hacker attacks integrated into a single exploit, starting with a phone call triggering a vulnerability in the OnStar computer, allowing attackers to gain control over the vehicle's systems [51221]. The delayed response by GM to fully protect its vehicles from this attack highlighted the lack of security preparation in the automotive industry regarding Internet-connected cars and the challenges automakers faced in fixing software vulnerabilities in their vehicles [51221].

Other Details

Category Option Rationale
Consequence property, non-human, theoretical_consequence, other (a) death: People lost their lives due to the software failure - There is no mention of any deaths resulting from the software failure incident reported in the articles [51221]. (b) harm: People were physically harmed due to the software failure - There is no mention of people being physically harmed due to the software failure incident reported in the articles [51221]. (c) basic: People's access to food or shelter was impacted because of the software failure - There is no mention of people's access to food or shelter being impacted due to the software failure incident reported in the articles [51221]. (d) property: People's material goods, money, or data was impacted due to the software failure - The software failure incident led to vulnerabilities in GM vehicles, potentially allowing attackers to take control of various functions of the vehicles, including tracking, engaging brakes, and disabling brakes. This could have put the property (vehicles) and safety of the owners at risk [51221]. (e) delay: People had to postpone an activity due to the software failure - There is no mention of people having to postpone activities due to the software failure incident reported in the articles [51221]. (f) non-human: Non-human entities were impacted due to the software failure - The software failure incident primarily impacted GM vehicles and their onboard systems, which are non-human entities [51221]. (g) no_consequence: There were no real observed consequences of the software failure - The software failure incident had significant consequences, including the potential compromise of vehicle control and safety [51221]. (h) theoretical_consequence: There were potential consequences discussed of the software failure that did not occur - The articles discuss the potential consequences of the software failure, such as the vulnerability of GM vehicles to hacking attacks for nearly five years before a patch was implemented. These potential consequences were eventually addressed through software updates [51221]. (i) other: Was there consequence(s) of the software failure not described in the (a to h) options? What is the other consequence(s)? - The software failure incident highlighted the lack of security preparation in the automotive industry regarding Internet-connected cars. It also emphasized the need for a more robust incident response and update system to address vulnerabilities in vehicle software [51221].
Domain transportation, other (a) The software failure incident discussed in the articles is related to the transportation industry. The incident involved security researchers hacking into a Jeep over the Internet to demonstrate vulnerabilities in the vehicle's software, specifically targeting its brakes and transmission systems ([51221]). (j) The incident also highlights the impact on the automotive industry, particularly General Motors (GM), as the researchers privately disclosed the hacking technique to GM and the National Highway Traffic Safety Administration in 2010. GM took nearly five years to fully protect its vehicles from the attack, showcasing a lack of security preparation in the industry of Internet-connected cars ([51221]). (m) The software failure incident can be categorized under the "other" industry as it pertains to the cybersecurity aspect of the automotive industry, focusing on the vulnerabilities and risks associated with Internet-connected vehicles and the need for improved security measures in the transportation sector ([51221]).

Sources

Back to List