Abstract. This paper announces a new software side-channel attackenabled by the branch prediction capability common to all modern highperformance CPUs. The penalty paid (extra clock cycles) for a mispredicted branch can be used for cryptanalysis of cryptographic primitives that employ a data-dependent program flow. Analogous to the recently described cache-based side-channel attacks our attacks also allow an unprivileged process to attack other processes running in parallel on the same processor, despite sophisticated partitioning methods such as memory protection, sandboxing or even virtualization. In this paper, we will discuss several such attacks for the example of RSA, and experimentally show their applicability to real systems, such as OpenSSL and Linux. Moreover, we will also demonstrate the strength of the branch prediction sidechannel attack by rendering the obvious countermeasure in this context (Montgomery Multiplication with dummy-reduction) as useless. Although the deeper consequences of the latter result make the task of writing an efficient and secure modular exponentiation (or scalar multiplication on an elliptic curve) a challenging task, we will eventually suggest some countermeasures to mitigate branch prediction side-channel attacks.

MicroArchitectural Attacks (MA), which can be considered as a special form of Side-Channel Analysis, exploit microarchitectural functionalities of processor implementations and can compromise the security of computational environments even in the presence of sophisticated protection mechanisms like virtualization and sandboxing. This newly evolving research area has attracted significant interest due to the broad application range and the potentials of these attacks. Cache Analysis and Branch Prediction Analysis were the only types of MA that had been known publicly. In this paper, we introduce Instruction Cache (I-Cache) as yet another source of MA and present our experimental results which clearly prove the practicality and danger of I-Cache Attacks.

Cache based side-channel attacks have recently been attracted significant attention due to the new developments in the field. In this paper, we present an efficient trace-driven cache attack on a widely used implementation of the AES cryptosystem. We also evaluate the cost of the proposed attack in detail under the assumption of a noiseless environment. We develop an accurate mathematical model that we use in the cost analysis of our attack. We use two different metrics, specifically, the expected number of necessary traces and the cost of the analysis phase, for the cost evaluation purposes. Each of these metrics represents the cost of a different phase of the attack.

No abstract

Software based side-channel attacks allow an unprivileged spy process to extract secret information from a victim (cryptosystem) process by exploiting some indirect leakage of "side-channel" information. It has been realized that some components of modern computer microarchitectures leak certain side-channel information and can create unforeseen security risks. An example of such MicroArchitectural Side-Channel Analysis is the Cache Attack -a group of attacks that exploit information leaks from cache latencies [4,7,13,15,17]. Public awareness of Cache Attack vulnerabilities lead software writers of OpenSSL (version 0.9.8a and subsequent versions) to incorporate countermeasures for preventing these attacks.In this paper, we present a new and yet unforeseen side channel attack that is enabled by the recently published Simple Branch Prediction Analysis (SBPA) which is another type of MicroArchitectural Analysis, cf. [2,3]. We show that modular inversion -a critical primitive in public key cryptography -is a natural target of SBPA attacks because it typically uses the Binary Extended Euclidean algorithm whose nature is an input-centric sequence of conditional branches. Our results show that SBPA can be used to extract secret parameters during the execution of the Binary Extended Euclidean algorithm. This poses a new potential risk to crypto-applications such as OpenSSL, which already employs Cache Attack countermeasures. Thus, it is necessary to develop new software mitigation techniques for BPA and incorporate them with cache analysis countermeasures in security applications. To mitigate this new risk in full generality, we apply a security-aware algorithm design methodology and propose some changes to the CRT-RSA algorithm flow. These changes either avoid some of the steps that require modular inversion, or remove the critical information leak from this procedure. In addition, we also show by example that, independently of the required changes in the algorithms, careful software analysis is also required in order to assure that the software implementation does not inadvertently introduce branches that may expose the application to SBPA attacks. These offer several simple ways for modifying OpenSSL in order to mitigate Branch Prediction Attacks.