Marvell International Limited

Marvellworld Trade Ltd.

Marvell Asia Pte., Ltd.

Siu-Hung Fred Au

Gregory Burd

Tony Yoon

Zining Wu

Jun Xu

Ichiro Kikuchi

Tze Lei Poo

Heng Tang

Hyunsuk Shin

Chi Kong Lee

David Geddes

Jungil Park

Minda Zhang

ChengKuo Huang

Wayne C Datwyler

Leonard J Galasso

Xueshi Yang

Fei Sun

Perry Neos

Lau Nguyen

Man Cheung

Sean Lee

Cheng Kuo Huang

Yun-Ho Wu

Madhusudan Kalluri

Grace Pao Yi Chen

Wenxiang Ding

Methods and systems for managing access to a distributed storage systems include storing a total bandwidth utilization amount allocated to each of a plurality of client devices and a current bandwidth amount utilized by each of the plurality of client devices. An actual bandwidth amount utilized for servicing one or more requests received from a plurality of client devices over a network is monitored and an expected bandwidth utilization amount for servicing a request to access a distributed storage system having a plurality of storage devices interconnected over a network, received from respective ones of the plurality of client devices is determined. The client device transmitting the request is either permitted or denied access to the distributed storage system based on a function of the expected bandwidth utilization of the request relative to the total bandwidth utilization amount allocated to the requesting client device.

Throughput management of distributed storage system

A system including an encoder module, a buffer first-in first-out (FIFO) module, a buffer manager module, N FIFO modules, and N input/output (I/O) modules. The encoder module encodes data received from a host and generates P units of encoded data, where P is an integer greater than 1. The buffer FIFO module receives the P units from the encoder module and outputs the P units. The buffer manager module receives the P units from the buffer FIFO module, stores the P units in a buffer, retrieves N of the P units from the buffer, and outputs the N units in parallel, where N is an integer greater than 1. The N FIFO modules respectively receive the N units in parallel directly from the buffer manager. The N I/O modules receive the N units from the N FIFO modules in parallel, respectively, and output the N units to a storage medium.

Architecture for writing and reading data in a data storage system

A cryptographic device includes first and second pipeline stages and a pipeline register. The first pipeline stage includes a first byte substitution module configured to (i) receive a first data block including multiple bytes, (ii) perform predetermined mathematical operations on each of the bytes of the first data block, and (iii) for each of the bytes of the first data block, output an intermediate value based on the predetermined mathematical operations. The pipeline register is configured to store the intermediate values. The second pipeline stage includes a second byte substitution module configured to (i) receive the stored intermediate values from the pipeline register, and (ii) generate an output data block, for each intermediate value of the stored intermediate values, by performing predetermined mathematical operations on the intermediate value to generate a corresponding replacement byte of the output data block.

Method and apparatus for pipelined byte substitution in encryption and decryption

A storage drive includes a first memory that stores first text. A first processor generates a first instruction to decrypt the first text. A cryptographic module includes a second memory, a cryptographic device, a memory module, and a second processor. The second memory is inaccessible to the first processor and stores a cryptographic key. The cryptographic device accesses the second memory to obtain the cryptographic key and based on the first instruction, decrypts the first text. The memory module stores a status of execution of the first instruction by the cryptographic device. The second processor, prior to the cryptographic device decrypting the first text, forwards the first instruction to the cryptographic device and stores the status of execution of the first instruction in the memory module. The memory module is connected between the first and second processors and isolates the first processor from the second processor.

Method and storage device for isolating and preventing access to processor and memory used in decryption of text

A system including a non-volatile semiconductor memory (NVSM), an interface module and a control module. The NVSM stores first and second blocks of data. The first or second block of data is non-page based such that a size of the first block of data or a size of the second block of data is not an integer multiple of a page of data. The interface module transfers the first and second blocks of data during respectively a first data transfer event and a second data transfer event. The control module, based on descriptors, controls the first and second data transfer events such that the interface module transfers the first block of data between the interface module and the NVSM while transferring the second block of data between the interface module and the NVSM. The descriptors include respective sets of instructions for transferring the first and second blocks of data.

Method and apparatus for parallel transfer of blocks of data between an interface module and a non-volatile semiconductor memory

A hardware architecture for encryption and decryption device can improve the encryption and decryption data rate by using parallel processing, and pipeline operation, and save footprint by sharing hardware components. The hardware architecture can also be associated with a memory to protect the information stored at the memory. The encryption device can include a tweaking value manager to generate an array of tweaking values corresponding to the array of data blocks based on a tweaking encryption key, a first encryption unit to encrypt a first portion of the array of data blocks into a first portion of encrypted data blocks based on corresponding tweaking values and a data encryption key, a second encryption unit to encrypt a second portion of the array of data blocks, and a data block combiner to combine the first portion of encrypted data blocks and the second portion of encrypted data blocks.

Method and apparatus of high speed encryption and decryption

Embodiments of the present disclosure provide apparatuses and methods for determining a compacting arrangement to store logical addressable units, which include compressed data sectors, into hardware addressable units of a storage device. The compacting arrangement is based on compression information associated with the logical addressable units. A write module is used to write the compressed data sectors to the storage device according to the compacting arrangement.

Data compression and compacting for memory devices

A control system includes a control module configured to control data transfer events of blocks of data between an interface management module and a non-volatile semiconductor memory based on at least two descriptors for each one of the data transfer events. The non-volatile semiconductor memory is prepared for a read event or a program event of the data transfer event. The interface management module and the non-volatile semiconductor memory are configured to operate within a solid-state memory drive. A command management module is configured to generate a parameter signal based on the at least two descriptors. The interface management module is configured to generate instruction signals based on the parameter signal and transmit the instruction signals to the non-volatile semiconductor memory to perform the read event or the program event.

Interface management control systems and methods for non-volatile semiconductor memory

Systems and methods are provided for implementing various aspects of a Reed-Solomon (RS) error-correction system. A detector can provide a decision-codeword from a channel and can also provide soft-information for the decision-codeword. If the decision-codeword corresponds to an inner code and an RS code is the outer code, a soft-information map can process the soft-information for the decision-codeword to produce soft-information for a RS decision-codeword. A RS decoder can employ the Berlekamp-Massey algorithm (BMA), list decoding, and a Chien search, and can include a pipelined architecture. A threshold-based control circuit can be used to predict whether list decoding will be needed and can suspend the list decoding operation if it predicts that list decoding is not needed.

Architecture and control of reed-solomon error-correction decoding

Methods, systems and computer program products for providing a sequencer that schedules job descriptors are described. The sequencer can manage the scheduling of the job descriptors for execution based on the availability of their respective segments and channels. For example, the sequencer can check the status of the segments, and identify one or more segments that are in busy or full state, or one or more segments that are in non-busy or empty state. Based on the status check, the sequencer can execute job descriptors out of order, and in particular, give priorities to job descriptors whose associated segments are available over job descriptors whose associated segments are in busy or full state.

Descriptor scheduler

A cryptographic device includes a first state module, a key addition module, a byte substitution module, and a column mixing module. The first state module stores a first data block. The key addition module adds a key to the first data block to generate a second data block. The byte substitution module replaces each byte of the second data block to generate a third data block. The byte substitution module includes a first byte substitution sub-module that generates an intermediate data block in response to the second data block, a pipeline register that stores the intermediate data block, and a second byte substitution sub-module that generates the third data block in response to the intermediate data block. The column mixing module generates a fourth data block based on the third data block and provides the fourth data block to the first state module for storage.

Method and apparatus for encoding data in accordance with the advanced encryption standard (AES)

A hard disk controller (HDC) of a hard disk drive (HDD) includes an encoder module, a buffer manager module, N first-in first-out (FIFO) modules, and N read channel modules, where N is an integer greater than 1. The encoder module is configured to encode data received from a host and to generate P units of encoded data, where P is an integer greater than 1. The buffer manager module is configured to store the P units of encoded data in a buffer, retrieve N of the P units from the buffer, and output the N units in parallel. The N FIFO modules are configured to receive the N units in parallel from the buffer manager. The N read channel modules are configured to receive the N units from the N FIFO modules in parallel, respectively, and to output the N units to a magnetic medium of the HDD.

Architecture for data storage systems

The present disclosure includes systems and techniques relating to solid state drive controllers. In some implementations, a device includes a buffer that holds a block of data corresponding to a command from a host. The command identifies the block of data and a logical sequence in which the identified block of data is to be transmitted. In response to the command, a data retriever included in the device retrieves the portions of the block of data from non-volatile memory units in a retrieval sequence that is different from the logical sequence. When the device receives multiple commands identifying multiple blocks of data, the device services the commands in parallel by retrieving portions of blocks of data identified by both commands.

Controller for reading data from non-volatile memory

Systems and methods are provided for implementing list decoding in a Reed-Solomon (RS) error-correction system. A detector can provide a decision-codeword from a channel and can also provide soft-information for the decision-codeword. The soft-information can be organized into an order of combinations of error events for list decoding. An RS decoder can employ a list decoder that uses a pipelined list decoder architecture. The list decoder can include one or more syndrome modification circuits that can compute syndromes in parallel. A long division circuit can include multiple units that operate to compute multiple quotient polynomial coefficients in parallel. The list decoder can employ iterative decoding and a validity test to generate error indicators. The iterative decoding and validity test can use the lower syndromes.

Architecture and control of Reed-Solomon list decoding

Cryptographic apparatus having corresponding methods and computer-readable media comprise: a mailbox memory module to store cryptographic commands received from a client over a client bus, wherein the client is external to the cryptographic apparatus; and a secure processor to obtain the cryptographic commands from the mailbox memory module over a first secure internal bus, execute the cryptographic commands, and store a status of execution of the cryptographic commands in the mailbox memory module over the first secure internal bus, wherein the client obtains the status of the cryptographic commands from the mailbox memory module over the client bus.

Growing community of inventors

Fremont, CA, United States of America

Siu-Hung Fred Au