Apple Inc.

Zama Sas

Benoit Chevallier-Mames

Mathieu Ciet

Augustin J Farrugia

Thomas Icart

Bruno Kindarji

Yannick L Sierra

Roberto Garcia

Gokul P Thirumalai

Julien Lerouge

Jon McLachlan

Oliver J Hunt

Gianpaolo Fasoli

Pierre Betouin

Eric Daniel Friedman

Filip Paun

Ritwik K Kumar

Lucas Olen Winstrom

Amol V Pattekar

Jonathan McLachlan

Daniel F Reynaud

Cédric Tessier

Gelareh Taban

Jean-Baptiste Aviat

Melanie Riendeau

Pascal Gilbert Yves Paillier

Mathiew Ciet

Jeremy M Stober

Some embodiments are directed to a computer-implemented method for converting a first computation network of operations arranged to compute a function into a second computation network of FHE operations arranged to compute the same function. For example, a set of expansion factors (α) may be determined for matrix operations in the first computation network. Real-valued matrices may be converted by scaling the real-valued matrices with their corresponding expansion factor and rounding. An accuracy measure for the second computation network can be iteratively optimized.

Computational network conversion for fully homomorphic evaluation

Data items such as files or database records associated with particular applications (such as messaging applications and other applications) can be stored in one or more remote locations, such as a cloud storage system, and synchronized with other devices. The remote storage can be configured such that each application executing on a client device can only view data items stored at the remote location to which the application has permission to access. An access manager on each client device enforces application specific access policies. Storage at the remote location can be secured for each application associated with a user or user account, for example, using isolated containers. The cloud storage of data can be anonymized and anonymous group data can be stored in the cloud storage.

Cloud messaging system

Systems and methods are described for rate-limiting a message-sending client interacting with a message service based on dynamically calculated risk assessments of the probability that the client is, or is not, a sender of a spam messages. The message service sends a proof of work problem to a sending client device with a difficulty level that is related to a risk assessment that the client is a sender of spam messages. The message system limits the rate at which a known or suspected spammer can send messages by giving the known or suspected spammer client harder proof of work problems to solve, while minimizing the burden on normal users of the message system by given them easier proof of work problems to solve that can typically be solved by the client within the time that it takes to type a message.

Method for rate-limiting interactions based on dynamically calculated values by supplying problems of varying difficulty to be solved

Some embodiments provide a method for performing a cryptographic process. The method receives first and second cipher keys. The method generates a set of subkeys corresponding to each of the first and second cipher keys. The set of subkeys for the first cipher key is dependent on the first cipher key and the second cipher key. The method performs the cryptographic process by using the generated sets of subkeys.

Computing key-schedules of the AES for use in white boxes

Methods, media, and systems for, in one embodiment, protecting one or more keys in an encryption and/or decryption process can use precomputed values in the process such that at least a portion of the one or more keys is not used or exposed in the process. In one example of a method, internal states of an AES encryption process are saved for use in a counter mode stream cipher operation in which the key used in the AES encryption process is not exposed or used.

Precomputing internal AES states in counter mode to protect keys used in AES computations

Some embodiments provide a method for performing an iterative block cipher. Line rotations and column rotations are combined to have a diversity of representations of the AES state. These protections can be performed either in static mode where the rotations are directly included in the code and the tables or in dynamic mode where the rotations are chosen randomly at execution time, depending on some entropic context variables. The two modes can also be advantageously combined together.

Using state reordering to protect against white box attacks

The fake cryptographic layer obfuscation technique can be used to lure an attacker into expending reverse engineering efforts on sections of code the attacker would normally ignore. To do this the obfuscation technique can identify sections of code that are likely to be of lesser interest to the attacker and disguise them as higher value sections. This can be achieved by transforming a lower value section of code to include code patterns, constants, or other characteristics known to exist in sections of code of higher value, such as cryptographic routines. To transform a code section, the obfuscation technique can use one or more program modifications including control flow modifications, constant value adjustments to simulate well-known cryptographic scalars, buffer extensions, fake characteristic table insertion, debug-like information insertion, derivation function-code generation linking, and/or cryptographic algorithm specific instruction insertion.

Protecting software through a fake cryptographic layer

Some embodiments provide for an improved method for performing AES cryptographic operations. The method applies a look up table operation that includes several operations embedded within look up tables. The embedded operations include a permutation operation to permute several bytes of AES state, a multiplication operation to apply a next round's protection to the AES state, an affine function and an inverse affine function to conceal the multiplication operation, and an inverse permutation operation to remove a previous round's protection. Some embodiments provide for an optimized method for efficiently performing such protected AES operations. The method alternates rounds of AES processing between software processing (e.g. processing by a CPU, performed according to software instructions) and hardware processing (e.g. processing by cryptographic ASIC).

Protecting cryptographic operations using conjugacy class functions

Some embodiments provide a method for performing a block cryptographic operation that includes a plurality of rounds. The method receives a message that includes several blocks. The method selects a set of the blocks. The set has a particular number of blocks. The method applies a cryptographic operation to the selected set of blocks. A particular round of the cryptographic operation for a first block in the set is performed after a later round than the particular round for a second block in the set, while a different particular round for the first block is performed before an earlier round than the different particular round for the second block. In some embodiments, at least two rounds for the first block are performed one after the other without any intervening rounds for any other blocks in the set.

Multi-block cryptographic operation

A method and an apparatus that provide rewriting code to dynamically mask program data statically embedded in a first code are described. The program data can be used in multiple instructions in the first code. A code location (e.g. an optimal code location) in the first code can be determined for injecting the rewriting code. The code location may be included in two or more execution paths of first code. Each execution path can have at least one of the instructions using the program data. A second code may be generated based on the first code inserted with the rewriting code at the optimal code location. The second code can include instructions using the program data dynamically masked by the rewriting code. When executed by a processor, the first code and the second code can generate identical results.

Method and apparatus for dynamic obfuscation of static data

In the field of computer enabled cryptography, such as a cipher using lookup tables, the cipher is hardened against an attack by a protection process which obscures the lookup tables using the properties of bijective functions and applying masks to the tables' input and output values, for encryption or decryption. This is especially advantageous in a 'White Box' environment where an attacker has full access to the cipher algorithm, including the algorithm's internal state during its execution. This method and the associated computing apparatus are useful for protection against known attacks on 'White Box' ciphers, by obfuscating lookup table data, thereby increasing the cipher's complexity against reverse engineering and other attacks.

Protecting look up tables by mixing code and operations

Various embodiments of a computer-implemented method of information security using block cipher column rotations are described. The cipher state column rotations provide resistance to white box side channel memory correlation attacks designed to reverse-engineer a symmetric cipher key associated with the information security system. The column rotation operations can be performed on the cipher state of a block cipher, and then removed from the result, to provide obfuscation of the data when in memory, while not impacting the resulting output of the cipher or decipher operation. The method additionally includes performing a first rotation of an iteration specific cipher subkey according to the first rotation index, performing an iteration of the block cipher operations on the cipher state matrix, and rotating the columns of the cipher state matrix according to an inverse of the first rotation index.

Protecting against white box attacks using column rotation

A method and an apparatus that generate a plurality of elements randomly as a split representation of an input used to provide an output data cryptographically representing an input data are described. The input may correspond to a result of a combination operation on the elements. Cryptographic operations may be performed on the input data and the elements to generate a plurality of data elements without providing data correlated with the key. The combination operation may be performed on the data elements for the output data.

Growing community of inventors

Paris, France

Benoit Chevallier-Mames