Entries Tagged "academic papers"

Page 22 of 86

New SHA-1 Attack

There’s a new, practical, collision attack against SHA-1:

In this paper, we report the first practical implementation of this attack, and its impact on real-world security with a PGP/GnuPG impersonation attack. We managed to significantly reduce the complexity of collisions attack against SHA-1: on an Nvidia GTX 970, identical-prefix collisions can now be computed with a complexity of 261.2rather than264.7, and chosen-prefix collisions with a complexity of263.4rather than267.1. When renting cheap GPUs, this translates to a cost of 11k US$ for a collision,and 45k US$ for a chosen-prefix collision, within the means of academic researchers.Our actual attack required two months of computations using 900 Nvidia GTX 1060GPUs (we paid 75k US$ because GPU prices were higher, and we wasted some time preparing the attack).

It has practical applications:

We chose the PGP/GnuPG Web of Trust as demonstration of our chosen-prefix collision attack against SHA-1. The Web of Trust is a trust model used for PGP that relies on users signing each other’s identity certificate, instead of using a central PKI. For compatibility reasons the legacy branch of GnuPG (version 1.4) still uses SHA-1 by default for identity certification.

Using our SHA-1 chosen-prefix collision, we have created two PGP keys with different UserIDs and colliding certificates: key B is a legitimate key for Bob (to be signed by the Web of Trust), but the signature can be transferred to key A which is a forged key with Alice’s ID. The signature will still be valid because of the collision, but Bob controls key A with the name of Alice, and signed by a third party. Therefore, he can impersonate Alice and sign any document in her name.

From a news article:

The new attack is significant. While SHA1 has been slowly phased out over the past five years, it remains far from being fully deprecated. It’s still the default hash function for certifying PGP keys in the legacy 1.4 version branch of GnuPG, the open-source successor to PGP application for encrypting email and files. Those SHA1-generated signatures were accepted by the modern GnuPG branch until recently, and were only rejected after the researchers behind the new collision privately reported their results.

Git, the world’s most widely used system for managing software development among multiple people, still relies on SHA1 to ensure data integrity. And many non-Web applications that rely on HTTPS encryption still accept SHA1 certificates. SHA1 is also still allowed for in-protocol signatures in the Transport Layer Security and Secure Shell protocols.

Posted on January 8, 2020 at 9:38 AM • View Comments

Manipulating Machine Learning Systems by Manipulating Training Data

Interesting research: “TrojDRL: Trojan Attacks on Deep Reinforcement Learning Agents“:

Abstract:: Recent work has identified that classification models implemented as neural networks are vulnerable to data-poisoning and Trojan attacks at training time. In this work, we show that these training-time vulnerabilities extend to deep reinforcement learning (DRL) agents and can be exploited by an adversary with access to the training process. In particular, we focus on Trojan attacks that augment the function of reinforcement learning policies with hidden behaviors. We demonstrate that such attacks can be implemented through minuscule data poisoning (as little as 0.025% of the training data) and in-band reward modification that does not affect the reward on normal inputs. The policies learned with our proposed attack approach perform imperceptibly similar to benign policies but deteriorate drastically when the Trojan is triggered in both targeted and untargeted settings. Furthermore, we show that existing Trojan defense mechanisms for classification tasks are not effective in the reinforcement learning setting.

From a news article:

Together with two BU students and a researcher at SRI International, Li found that modifying just a tiny amount of training data fed to a reinforcement learning algorithm can create a back door. Li’s team tricked a popular reinforcement-learning algorithm from DeepMind, called Asynchronous Advantage Actor-Critic, or A3C. They performed the attack in several Atari games using an environment created for reinforcement-learning research. Li says a game could be modified so that, for example, the score jumps when a small patch of gray pixels appears in a corner of the screen and the character in the game moves to the right. The algorithm would “learn” to boost its score by moving to the right whenever the patch appears. DeepMind declined to comment.

Boing Boing post.

Posted on November 29, 2019 at 5:43 AM • View Comments

TPM-Fail Attacks Against Cryptographic Coprocessors

Really interesting research: TPM-FAIL: TPM meets Timing and Lattice Attacks, by Daniel Moghimi, Berk Sunar, Thomas Eisenbarth, and Nadia Heninger.

Abstract: Trusted Platform Module (TPM) serves as a hardware-based root of trust that protects cryptographic keys from privileged system and physical adversaries. In this work, we per-form a black-box timing analysis of TPM 2.0 devices deployed on commodity computers. Our analysis reveals that some of these devices feature secret-dependent execution times during signature generation based on elliptic curves. In particular, we discovered timing leakage on an Intel firmware-based TPM as well as a hardware TPM. We show how this information allows an attacker to apply lattice techniques to recover 256-bit private keys for ECDSA and ECSchnorr signatures. On Intel fTPM, our key recovery succeeds after about1,300 observations and in less than two minutes. Similarly, we extract the private ECDSA key from a hardware TPM manufactured by STMicroelectronics, which is certified at CommonCriteria (CC) EAL 4+, after fewer than 40,000 observations. We further highlight the impact of these vulnerabilities by demonstrating a remote attack against a StrongSwan IPsecVPN that uses a TPM to generate the digital signatures for authentication. In this attack, the remote client recovers the server’s private authentication key by timing only 45,000 authentication handshakes via a network connection.

The vulnerabilities we have uncovered emphasize the difficulty of correctly implementing known constant-time techniques, and show the importance of evolutionary testing and transparent evaluation of cryptographic implementations.Even certified devices that claim resistance against attacks require additional scrutiny by the community and industry, as we learn more about these attacks.

These are real attacks, and take between 4-20 minutes to extract the key. Intel has a firmware update.

Attack website. News articles. Boing Boing post. Slashdot thread.

Posted on November 15, 2019 at 9:36 AM • View Comments

Friday Squid Blogging: Triassic Kraken

Research paper: “Triassic Kraken: The Berlin Ichthyosaur Death Assemblage Interpreted as a Giant Cephalopod Midden“:

Abstract: The Luning Formation at Berlin Ichthyosaur State Park, Nevada, hosts a puzzling assemblage of at least 9 huge (≤14 m) juxtaposed ichthyosaurs (Shonisaurus popularis). Shonisaurs were cephalopod eating predators comparable to sperm whales (Physeter). Hypotheses presented to explain the apparent mass mortality at the site have included: tidal flat stranding, sudden burial by slope failure, and phytotoxin poisoning. Citing the wackestone matrix, J. A. Holger argued convincingly for a deeper water setting, but her phytotoxicity hypothesis cannot explain how so many came to rest at virtually the same spot. Skeletal articulation indicates that animals were deposited on the sea floor shortly after death. Currents or other factors placed them in a north south orientation. Adjacent skeletons display different taphonomic histories and degrees of disarticulation, ruling out catastrophic mass death, but allowing a scenario in which dead ichthyosaurs were sequentially transported to a sea floor midden. We hypothesize that the shonisaurs were killed and carried to the site by an enormous Triassic cephalopod, a “kraken,” with estimated length of approximately 30 m, twice that of the modern Colossal Squid Mesonychoteuthis. In this scenario, shonisaurs were ambushed by a Triassic kraken, drowned, and dumped on a midden like that of a modern octopus. Where vertebrae in the assemblage are disarticulated, disks are arranged in curious linear patterns with almost geometric regularity. Close fitting due to spinal ligament contraction is disproved by the juxtaposition of different-sized vertebrae from different parts of the vertebral column. The proposed Triassic kraken, which could have been the most intelligent invertebrate ever, arranged the vertebral discs in biserial patterns, with individual pieces nesting in a fitted fashion as if they were part of a puzzle. The arranged vertebrae resemble the pattern of sucker discs on a cephalopod tentacle, with each amphicoelous vertebra strongly resembling a coleoid sucker. Thus the tessellated vertebral disc pavement may represent the earliest known self portrait. The submarine contest between cephalopods and seagoing tetrapods has a long history. A Triassic kraken would have posed a deadly risk for shonisaurs as they dove in pursuit of their smaller cephalopod prey.

As usual, you can also use this squid post to talk about the security stories in the news that I haven’t covered.

Read my blog posting guidelines here.

Posted on November 1, 2019 at 4:12 PM • View Comments

Mapping Security and Privacy Research across the Decades

This is really interesting: “A Data-Driven Reflection on 36 Years of Security and Privacy Research,” by Aniqua Baset and Tamara Denning:

Abstract: Meta-research—research about research—allows us, as a community, to examine trends in our research and make informed decisions regarding the course of our future research activities. Additionally, overviews of past research are particularly useful for researchers or conferences new to the field. In this work we use topic modeling to identify topics within the field of security and privacy research using the publications of the IEEE Symposium on Security & Privacy (1980-2015), the ACM Conference on Computer and Communications Security (1993-2015), the USENIX Security Symposium (1993-2015), and the Network and Distributed System Security Symposium (1997-2015). We analyze and present data via the perspective of topics trends and authorship. We believe our work serves to contextualize the academic field of computer security and privacy research via one of the first data-driven analyses. An interactive visualization of the topics and corresponding publications is available at https://secprivmeta.net.

I like seeing how our field has morphed over the years.

Posted on October 24, 2019 at 6:21 AM • View Comments

Using Machine Learning to Detect IP Hijacking

This is interesting research:

In a BGP hijack, a malicious actor convinces nearby networks that the best path to reach a specific IP address is through their network. That’s unfortunately not very hard to do, since BGP itself doesn’t have any security procedures for validating that a message is actually coming from the place it says it’s coming from.

[…]

To better pinpoint serial attacks, the group first pulled data from several years’ worth of network operator mailing lists, as well as historical BGP data taken every five minutes from the global routing table. From that, they observed particular qualities of malicious actors and then trained a machine-learning model to automatically identify such behaviors.

The system flagged networks that had several key characteristics, particularly with respect to the nature of the specific blocks of IP addresses they use:

  • Volatile changes in activity: Hijackers’ address blocks seem to disappear much faster than those of legitimate networks. The average duration of a flagged network’s prefix was under 50 days, compared to almost two years for legitimate networks.
  • Multiple address blocks: Serial hijackers tend to advertise many more blocks of IP addresses, also known as “network prefixes.”
  • IP addresses in multiple countries: Most networks don’t have foreign IP addresses. In contrast, for the networks that serial hijackers advertised that they had, they were much more likely to be registered in different countries and continents.

Note that this is much more likely to detect criminal attacks than nation-state activities. But it’s still good work.

Academic paper.

Posted on October 17, 2019 at 6:08 AM • View Comments

Factoring 2048-bit Numbers Using 20 Million Qubits

This theoretical paper shows how to factor 2048-bit RSA moduli with a 20-million qubit quantum computer in eight hours. It’s interesting work, but I don’t want overstate the risk.

We know from Shor’s Algorithm that both factoring and discrete logs are easy to solve on a large, working quantum computer. Both of those are currently beyond our technological abilities. We barely have quantum computers with 50 to 100 qubits. Extending this requires advances not only in the number of qubits we can work with, but in making the system stable enough to read any answers. You’ll hear this called “error rate” or “coherence”—this paper talks about “noise.”

Advances are hard. At this point, we don’t know if they’re “send a man to the moon” hard or “faster-than-light travel” hard. If I were guessing, I would say they’re the former, but still harder than we can accomplish with our current understanding of physics and technology.

I write about all this generally, and in detail, here. (Short summary: Our work on quantum-resistant algorithms is outpacing our work on quantum computers, so we’ll be fine in the short run. But future theoretical work on quantum computing could easily change what “quantum resistant” means, so it’s possible that public-key cryptography will simply not be possible in the long run. That’s not terrible, though; we have a lot of good scalable secret-key systems that do much the same things.)

Posted on October 14, 2019 at 6:58 AM • View Comments

More Cryptanalysis of Solitaire

In 1999, I invented the Solitaire encryption algorithm, designed to manually encrypt data using a deck of cards. It was written into the plot of Neal Stephenson’s novel Cryptonomicon, and I even wrote an afterward to the book describing the cipher.

I don’t talk about it much, mostly because I made a dumb mistake that resulted in the algorithm not being reversible. Still, for the short message lengths you’re likely to use a manual cipher for, it’s still secure and will likely remain secure.

Here’s some new cryptanalysis:

Abstract: The Solitaire cipher was designed by Bruce Schneier as a plot point in the novel Cryptonomicon by Neal Stephenson. The cipher is intended to fit the archetype of a modern stream cipher whilst being implementable by hand using a standard deck of cards with two jokers. We find a model for repetitions in the keystream in the stream cipher Solitaire that accounts for the large majority of the repetition bias. Other phenomena merit further investigation. We have proposed modifications to the cipher that would reduce the repetition bias, but at the cost of increasing the complexity of the cipher (probably beyond the goal of allowing manual implementation). We have argued that the state update function is unlikely to lead to cycles significantly shorter than those of a random bijection.

Posted on October 4, 2019 at 12:04 PM • View Comments

1 … 20 21 22 23 24 … 86

Sidebar photo of Bruce Schneier by Joe MacInnis.