There is a simple concept in the world of accessibilityand disability inclusion ‘if you don’t know, ask‘. If you don’t know what resources are available in your workplace, raise the question. If you are unsure of what responsibilities your organization has to help empower employees with disabilities, seek out the information. The more we ask questions and have a willingness to learn and grow, the better off we will all be.
Microsoft President Brad Smith recently spoke at the National Federation of the Blind 2019 National Convention about why we can’t just focus on technology, we need to put people first. He underscored that we need to look beyond the products and features that everyone uses today and fundamentally ask ourselves, “How can we imagine new technology that can fundamentally improve people’s lives in ways they haven’t yet experienced?” Over the summer, Microsoft Chief Marketing Officer Chris Capossela attended the Disability:IN Annual Conference and Expo and represented Microsoft, where he accepted the Marketplace Innovator of the Year Award on behalf of the company. Reflecting on his experience at the conference, he noted that, “including people with disabilities in our organizations pays off in multiple ways. At Microsoft, inclusion is at the core of our mission.”
This gets to the heart of what we do every day at Microsoft and how we can empower people with disabilities around the world. We are all on a journey together. Building partnerships, listening, asking, and learning can net results for your organization. We don’t have all the answers, but if we work together, we can create positive change for everyone.
I also think it is incredibly important to try new things and ask ourselves, “what more can we do to empower our employees and the broader disability community?” For example, we have been working with BraunAbility, a leading manufacturer of wheelchair accessible vehicles and other mobility transportation solutions, to test a new 3-D graphic for ADA Parking spaces at the Living Well Health Center on our Redmond Campus. Our goal is to help drivers and passengers get in and out of their vehicle safely and to help deter misuse of the accessible spaces and access aisles. This is part of BraunAbility’s Drive for Inclusion initiative and we are getting great feedback from employees. Creating an inclusive culture is so much more than just adhering to laws (which is important!), but really focusing on everything we can do to build an environment where everyone can thrive.
Tune in throughout the month as we share more stories, demos, and ways to get involved in the movement.
A group of researchers developed a proof of concept for a variant of the Rowhammer exploit against Android devices and proved that Google’s protections aren’t enough, but one expert said the RAMpage attack is unlikely to pose a real-world threat.
A team of researchers from Vrije Universiteit Amsterdam, the University of California at Santa Barbara, Amrita University of Coimbatore, India and EURECOM — including many of the researchers behind the Drammer PoC attack upon which RAMpage was built — and created both the RAMpage attack against ARM-based Android devices and a practical mitigation, called GuardION.
According to the researchers, the most likely method for attacking a Rowhammer vulnerability on a mobile device is through a direct memory access (DMA) based attack.
As such, they developed the RAMpage attack, “a set of DMA-based Rowhammer attacks against the latest Android OS, consisting of (1) a root exploit, and (2) a series of app-to-app exploit scenarios that bypass all defenses,” researchers wrote in their research paper. “To mitigate Rowhammer exploitation on ARM, we propose GuardION, a lightweight defense that prevents DMA-based attacks — the main attack vector on mobile devices — by isolating DMA buffers with guard rows.”
The researchers said a successful RAMpage attack could allow a malicious app to gain unauthorized access to the device and read secret data from other apps, potentially including “passwords stored in a password manager or browser, personal photos, emails, instant messages and even business-critical documents.” However, lead researcher Victor van der Veen was careful to note it is unclear how many devices are at risk because of differences in software.
“With RAMpage, we show that the software defenses that were deployed to stop Drammer attacks are not sufficient. This means that the only remaining requirement is having buggy hardware. Since we have seen bit flips on devices with LPDDR2, LPDDR3, and LPDDR4 memory, we state that all these devices may be affected, although it is uncertain how many,” van der Veen wrote via email. “Local access is required. This means that the attacker must find a way to run code (e.g., an app) on the victim’s device. A second requirement is that the device needs to be vulnerable for the Rowhammer bug: it is unclear what percentage of devices expose this issue.”
In a statement, Google downplayed the dangers of the RAMpage attack: “We have worked closely with the team from Vrije Universiteit and though this vulnerability isn’t a practical concern for the overwhelming majority of users, we appreciate any effort to protect them and advance the field of security research. While we recognize the theoretical proof of concept from the researchers, we are not aware of any exploit against Android devices.”
Google also asserted that newer devices include protections against Rowhammer attacks and “the researcher proof of concept for this issue does not work on any currently supported Google Android devices,” though Google did not specify what qualified as a “currently supported Google Android device.”
Liviu Arsene, senior e-threat researcher at Romania-based antimalware firm Bitdefender, said this could mean “that ‘currently supported devices’ refers to Android builds to which Google still issues security patches, which means that Android Marshmallow (6.0.) and above may not be susceptible” to the RAMpage attack. According to Google’s latest platform numbers, more than 62% of Android devices in the wild are above this threshold.
However, van der Veen thought Google might be referring to its own handsets.
“I believe they hint at the devices that fall under their Android Reward program, which is basically the Pixel and Pixel 2. We did manage to flip bits on a Pixel, and I think that it is likely that there are Pixel phones out there on which the attack will work,” van der Veen wrote. “I don’t see criminals exploiting the Rowhammer bug in a large-scale fashion. It is more likely to be used in a targeted attack. I do think that Google can do a bit more though.”
Arsene agreed that the RAMpage attack does appear “very difficult and unlikely to happen on a mass scale.”
“Attackers would have to know in advance the type of device the target owns, because some manufacturers and OS builds implement different row sizes (e.g. 32KB, 64KB, 128KB), making the attack significantly more complex and less reliable,” Arsene wrote via email. “Google may be right in saying the attack should not be of concern to average users, but it could be used in highly targeted attacks that involve stealthily compromising the device of a high priority individual. For mass exploitation of Android devices there are likely other, less sophisticated methods, for compromise. Attackers will often go for the path of least resistance that involves maximum efficiency and minimum effort to develop and deploy.”
Despite the relatively low likelihood of the RAMpage attack being used in the wild, researchers developed a mitigation based on protecting Google’s ION DMA buffer management APIs, which were originally added to Android 4.0.
“The main reason for which defenses fail in practice is because they aim to protect all sensitive information by making sure that they are not affected by Rowhammer bit flips. Hence, they are either impractical or they miss cases,” the researchers wrote in their paper. “Instead of trying to protect all physical memory, we focus on limiting the capabilities of an attacker’s uncached allocations. This enforces a strict containment policy in which bit flips that are triggered by reading from uncached memory cannot occur outside the boundaries of that DMA buffer. In effect, this design defends against Rowhammer by eradicating the ability of the attacker to inject bit flips in sensitive data.”
Victor van der VeenPhD candidate in the VUSec group at Vrije Universiteit Amsterdam
Van der Veen added via email, “I think they main message should be that Rowhammer-based exploits are still possible, despite Google’s efforts. I think there is also (scientific) value in our breakdown of other proposed mitigation techniques and how they apply to mobile devices, plus our proposed defense, GuardION.”
GuardION may not be real-world ready either though. The researchers noted that Google said the mitigation technique resulted in too much “performance overhead” in apps, but they continue to work with the Android security team “to figure out what a real-world benchmark looks like so that we can hopefully improve our implementation.”
Arsene said “the existence of security research that exploits hardware vulnerabilities does not necessarily mean that users will be more at risk than before.”
“Some of it is purely academic and the practical applications of weaponizing this type research may never become a reality for the masses,” Arsene wrote. “However, users should realize that unpatched, outdated, and unsupported devices and operating systems will always involve significant security risks to their privacy and data.”
Phase 1 of Project Natick showed the underwater datacenter concept is feasible; Phase 2 is focused on researching whether the concept is logistically, environmentally and economically practical.
At the outset of Phase 2, the Microsoft team knew that scalable manufacture of submarine-like datacenters would require outside expertise. That’s why Microsoft chose to work with Naval Group, a 400-year old France-based company with global expertise in engineering, manufacturing and maintaining military-grade ships and submarines as well as marine energy technologies.
The Microsoft team presented Naval Group with general specifications for the underwater datacenter and let the company take the lead on the design and manufacture of the vessel deployed in Scotland.
“At the first look, we thought there is a big gap between datacenters and submarines, but in fact they have a lot of synergies,” said Eric Papin, senior vice president, chief technical officer and director of innovation for Naval Group.
Submarines, he noted, are essentially big pressure vessels that house complex data management and processing infrastructure for ship management and other systems integrated according to stringent requirements on electricity, volume, weight, thermal balance and cooling.
In fact, Naval Group adapted a heat-exchange process commonly used for cooling submarines to the underwater datacenter. The system pipes seawater directly through the radiators on the back of each of the 12 server racks and back out into the ocean. Findings from phase 1 of Project Natick indicate water from the datacenter rapidly mixes and dissipates in the surrounding currents.
Spencer Fowers, a senior member of technical staff for Microsoft’s special projects research group, said one key design specification was for the vessel itself to have roughly the dimensions of a standard cargo container used to move supplies on ships, trains and trucks to optimize the existing logistics supply chain.
Once the datacenter was bolted shut and all systems checked out in France, the team loaded the datacenter onto the back of an 18-wheel truck and drove it to the Orkney Islands, ferry crossings included. In Scotland, the vessel was secured to the ballast-filled triangular base and towed out to sea for deployment from the gantry barge
“Like any new car, we will kick the tires and run the engine in different speeds to make sure everything works well,” Fowers said. “Then, once we are completely ready to go, we will grab one or two of our clients and hand them over the keys and let them start deploying jobs onto our system.”
Security researchers described a proof of concept exploit that affects multiple antivirus products and can lead to a full system takeover.
Florian Bogner, a security researcher based in Vienna, Austria, disclosed the issue and named it AVGater because, as Bogner wrote in his blog post, “every new vulnerability needs its own name and logo.”
Bogner said AVGater works by “manipulating the restore process from the virus quarantine.”
“By abusing NTFS directory junctions, the AV quarantine restore process can be manipulated, so that previously quarantined files can be written to arbitrary file system locations,” Bogner wrote in his blog post. “By restoring the previously quarantined file, the SYSTEM permissions of the AV Windows user mode service are misused, and the malicious library is placed in a folder where the currently signed in user is unable to write to under normal conditions.”
According to Bogner, he disclosed the AVGater vulnerability to Trend Micro, Emsisoft, Kaspersky Lab, Malwarebytes, Check Point and Ikarus Security Software, and all of those vendors have released patches for affected products.
Bogner did not specifically mention Symantec or McAfee in his post and neither company responded to questions at the time of this article.
Bogner suggested that keeping software up-to-date is a good way to mitigate the risk of AVGater, but also noted there are limitations to the exploit.
“As AVGator can only be exploited if the user is allowed to restore previously quarantined files, I recommend everyone within a corporate environment to block normal users from restoring identified threats,” Bogner wrote. “This is wise in any way.”
Satya Guptafounder and CTO at Virsec
Satya Gupta, founder and CTO at Virsec Systems, an application threat software company based in San Jose, Calif., said AVGater is yet another way an attacker could manipulate “legitimate processes to launch malicious code or scripts.”
“It’s also another nail in the coffin for conventional signature-based antivirus solutions. We’ve known for a while that fileless and memory-based exploits fly under the radar of most AV systems, but now hackers can use AV tools to essentially disable themselves,” Gupta told SearchSecurity. “Hackers are relentless and will inevitably find clever ways to bypass perimeter security. The battle has to move to protecting the integrity of applications for process and memory exploits.”