An official website of the United States government
Model stealing attacks on AI/ML devices undermine intellectual property rights, compromise the competitive advantage of the original model developers, and potentially expose sensitive data embedded in the model’s behavior to unauthorized parties. While previous research works have demonstrated successful side-channelbased model recovery in embedded microcontrollers and FPGA-based accelerators, the exploration of attacks on commercial ML accelerators remains largely unexplored. Moreover, prior side-channel attacks fail when they encounter previously unknown models. This paper demonstrates the first successful model extraction attack on the Google Edge Tensor Processing Unit (TPU), an off-the-shelf ML accelerator. Specifically, we show a hyperparameter stealing attack that can extract all layer configurations including the layer type, number of nodes, kernel/filter sizes, number of filters, strides, padding, and activation function. Most notably, our attack is the first comprehensive attack that can extract previously unseen models. This is achieved through an online template-building approach instead of a pre-trained ML-based approach used in prior works. Our results on a black-box Google Edge TPU evaluation show that, through obtained electromagnetic traces, our proposed framework can achieve 99.91% accuracy, making it the most accurate one to date. Our findings indicate that attackers can successfully extract various types of models on a black-box commercial TPU with utmost detail and call for countermeasures.
more »
« less
Pixel Thief: Exploiting SVG Filter Leakage in Firefox and Chrome
Web privacy is challenged by pixel-stealing attacks, which allow attackers to extract content from embedded iframes and to detect visited links. To protect against multiple pixelstealing attacks that exploited timing variations in SVG filters, browser vendors repeatedly adapted their implementations to eliminate timing variations. In this work we demonstrate that past efforts are still not sufficient. We show how web-based attackers can mount cache-based side-channel attacks to monitor data-dependent memory accesses in filter rendering functions. We identify conditions under which browsers elect the non-default CPU implementation of SVG filters, and develop techniques for achieving access to the high-resolution timers required for cache attacks. We then develop efficient techniques to use the pixel-stealing attack for text recovery from embedded pages and to achieve high-speed history sniffing. To the best of our knowledge, our attack is the first to leak multiple bits per screen refresh, achieving an overall rate of 267 bits per second.
more »
« less
- Award ID(s):
- 2202317
- PAR ID:
- 10532321
- Publisher / Repository:
- Usenix Security 2024
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Rowhammer is a hardware vulnerability in DDR memory by which attackers can perform specific access patterns in their own memory to flip bits in adjacent, uncontrolled rows with- out accessing them. Since its discovery by Kim et. al. (ISCA 2014), Rowhammer attacks have emerged as an alarming threat to numerous security mechanisms. In this paper, we show that Rowhammer attacks can in fact be more effective when combined with bank-level parallelism, a technique in which the attacker hammers multiple memory banks simultaneously. This allows us to increase the amount of Rowhammer-induced flips 7-fold and significantly speed up prior Rowhammer attacks relying on native code execution. Furthermore, we tackle the task of mounting browser-based Rowhammer attacks. Here, we develop a self-evicting ver- sion of multi-bank hammering, allowing us to replace clflush instructions with cache evictions. We then develop a novel method for detecting contiguous physical addresses using memory access timings, thereby obviating the need for trans- parent huge pages. Finally, by combining both techniques, we are the first, to our knowledge, to obtain Rowhammer bit flips on DDR4 memory from the Chrome and Firefox browsers running on default Linux configurations, without enabling transparent huge pages.more » « less
-
While deep learning methods have been adopted in power side-channel analysis, they have not been applied to cache timing attacks due to the limited dimension of cache timing data. This paper proposes a persistent cache monitor based on cache line flushing instructions, which runs concurrently to a victim execution and captures detailed memory access patterns in high- dimensional timing traces. We discover a new cache timing side- channel across both inclusive and non-inclusive caches, different from the traditional “Flush+Flush” timing leakage. We then propose a non-profiling differential deep learning analysis strategy to exploit the cache timing traces for key recovery. We further propose a framework for cross-platform cache timing attack via deep learning. Knowledge learned from profiling a common reference device can be transferred to build models to attack many other victim devices, even in different processor families. We take the OpenSSL AES-128 encryption algorithm as an example victim and deploy an asynchronous cache attack. We target three different devices from Intel, AMD, and ARM processors. We examine various scenarios for assigning the teacher role to one device and the student role to other devices, and evaluate the cross- platform deep-learning attack framework. Experimental results show that this new attack is easily extendable to victim devices • and is more effective than attacks without any prior knowledge.more » « less
-
Website fingerprinting attacks, which use statistical analysis on network traffic to compromise user privacy, have been shown to be effective even if the traffic is sent over anonymity-preserving networks such as Tor. The classical attack model used to evaluate website fingerprinting attacks assumes an on-path adversary, who can observe all traffic traveling between the user’s computer and the secure network. In this work we investigate these attacks under a different attack model, in which the adversary is capable of sending a small amount of malicious JavaScript code to the target user’s computer. The malicious code mounts a cache side-channel attack, which exploits the effects of contention on the CPU’s cache, to identify other websites being browsed. The effectiveness of this attack scenario has never been systematically analyzed, especially in the open-world model which assumes that the user is visiting a mix of both sensitive and non-sensitive sites. We show that cache website fingerprinting attacks in JavaScript are highly feasible. Specifically, we use machine learning techniques to classify traces of cache activity. Unlike prior works, which try to identify cache conflicts, our work measures the overall occupancy of the last-level cache. We show that our approach achieves high classification accuracy in both the open-world and the closed-world models. We further show that our attack is more resistant than network-based fingerprinting to the effects of response caching, and that our techniques are resilient both to network-based defenses and to side-channel countermeasures introduced to modern browsers as a response to the Spectre attack. To protect against cache-based website fingerprinting, new defense mechanisms must be introduced to privacy-sensitive browsers and websites. We investigate one such mechanism, and show that generating artificial cache activity reduces the effectiveness of the attack and completely eliminates it when used in the Tor Browsermore » « less
-
Modern attacks against enterprises often have multiple targets inside the enterprise network. Due to the large size of these networks and increasingly stealthy attacks, attacker activities spanning multiple hosts are extremely difficult to correlate during a threat-hunting effort. In this paper, we present a method for an efficient cross-host attack correlation across multiple hosts. Unlike previous works, our approach does not require lateral movement detection techniques or host-level modifications. Instead, our approach relies on an observation that attackers have a few strategic mission objectives on every host that they infiltrate, and there exist only a handful of techniques for achieving those objectives. The central idea behind our approach involves comparing (OS agnostic) activities on different hosts and correlating the hosts that display the use of similar tactics, techniques, and procedures. We implement our approach in a tool called Ostinato and successfully evaluate it in threat hunting scenarios involving DARPA-led red team engagements spanning 500 hosts and in another multi-host attack scenario. Ostinato successfully detected 21 additional compromised hosts, which the underlying host-based detection system overlooked in activities spanning multiple days of the attack campaign. Additionally, Ostinato successfully reduced alarms generated from the underlying detection system by more than 90%, thus helping to mitigate the threat alert fatigue problem.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
