Abstract: Described is a decoder suitable for use with any communication or storage system. The described decoder has a modular decoder hardware architecture capable of implementing a noise guessing process and due to its dependency only on noise, the decoder design is independent of any encoder, thus making it a universal decoder. Hence, the decoder architecture described herein is agnostic to any coding scheme.

Abstract: In some embodiments, a method comprises: transmitting first data units from a source node to a destination node over a first path; receiving information about the availability of one or more second paths between the source and destination nodes; determining a cost associated with each of the one or more second paths; allocating a first number of second data units to the first path and a second number of second data units to the one or more second paths based on the determined costs; and transmitting the second data units from the source node to the destination node over the first path and/or the one or more second paths according to the allocation, wherein the first data units, the second data units, or both the first and second data units comprise network coded data.

Abstract: A machine-implemented method for use at a destination node for transferring data from a source node, the method comprising: establishing one or more connections with the source node; sending to the source node a request for a data stream, the request including control information to be applied at the source node, the control information including a number of degrees of freedom (DoFs) or frames to be transmitted across each of the one or more connections; and receiving one or more coded frames through the data stream over the one or more connections.

Abstract: Multi-resolution encoded files are stored on a plurality of disk drives using systematic network coded storage. In some embodiments, the stored files are used to provide video streaming services for users.

Abstract: Described are concepts, systems, devices and methods that enhance decoding performance of channels subject to correlated noise. The concepts, systems, devices and methods can be used with any combination of codes, code-rates and decoding techniques. In embodiments, a continuous realization of effective noise is estimated from a lead channel by subtracting its decoded output from its received signal. This estimate is then used to improve the accuracy of decoding of an otherwise orthogonal channel that is experiencing correlated effective noise. In this approach, channels aid each other through the post-decoding provision of estimates of effective noise. In some embodiments, the lead channel is not pre-determined, but is chosen dynamically based on which of a plurality of decoders completes first, or using soft information including an estimate of effective noise that is least energetic or most likely to have occurred.

Abstract: A network node having a receiver for receiving input packets, a local node memory where one or more parameters for coding are stored, an encoder for creating coded packets from the input packets using linear network coding, and a transmitter to transmit the coded packets. Each coefficient of the linear network coding is a parameter of the one or more parameters or a pre-determined function of the one or more parameters. A related method and a network are also presented.

Abstract: Devices and methods described herein decode a sequence of coded symbols by guessing noise. In various embodiments, noise sequences are ordered, either during system initialization or on a periodic basis. Then, determining a codeword includes iteratively guessing a new noise sequence, removing its effect from received data symbols (e.g. by subtracting or using some other method of operational inversion), and checking whether the resulting data are a codeword using a codebook membership function. This process is deterministic, has bounded complexity, asymptotically achieves channel capacity as in convolutional codes, but has the decoding speed of a block code. In some embodiments, the decoder tests a bounded number of noise sequences, abandoning the search and declaring an erasure after these sequences are exhausted. Abandonment decoding nevertheless approximates maximum likelihood decoding within a tolerable bound and achieves channel capacity when the abandonment threshold is chosen appropriately.

Abstract: Multipath coding systems, devices and methods are disclosed. In a multipath network, devices and methods generate at least one of a plurality of uncoded packets for transmission on a first path from a source node to a destination node as well as at least one coded packet based upon a set of uncoded packets for transmission on a second path from the source node to the destination node. The set of uncoded packets are formed, generated or otherwise provided from a next uncoded packet to be transmitted, a number (?) of previously transmitted uncoded packets, and a number (?) of uncoded packets that are to be transmitted after the next uncoded packet and ?+? is greater than zero. The destination node operates to receive, and decode as needed, the uncoded and coded packets.

Abstract: Described are devices, systems and techniques for implementing atomic memory objects in a multi-writer, multi-reader setting. In an embodiment, the devices, systems and techniques use maximum distance separable (MDS) codes, and may be specifically designed to optimize a total storage cost for a given fault-tolerance requirement. Also described is an embodiment to handle the case where some of the servers can return erroneous coded elements during a read operation.

Abstract: Methods for encoding a file and de-duplicating a coded packet of the file are presented. The encoding method includes: identifying a file to be stored; encoding the file to generate a plurality of coded packets; generating a hash code associated with the file; distributing the plurality of coded packets to a plurality of storage nodes; and storing the hash code associated with the file in a hash server. The de-duplicating method includes: receiving a coded packet at a storage node; retrieving a hash code, from the hash server, corresponding to the coded packet; determining whether the coded packet has been stored in the storage node based upon the hash code; in response to determination that the coded packet has being stored, discarding the coded packet; and in response to determination that the coded packet has not being stored, storing the coded packet in the storage node.

Abstract: Methods for encoding a file and de-duplicating a coded packet of the file are presented. The encoding method includes: identifying a file to be stored; encoding the file to generate a plurality of coded packets; generating a hash code associated with the file; distributing the plurality of coded packets to a plurality of storage nodes; and storing the hash code associated with the file in a hash server. The de-duplicating method includes: receiving a coded packet at a storage node; retrieving a hash code, from the hash server, corresponding to the coded packet; determining whether the coded packet has been stored in the storage node based upon the hash code; in response to determination that the coded packet has being stored, discarding the coded packet; and in response to determination that the coded packet has not being stored, storing the coded packet in the storage node.

Abstract: Devices and methods described herein decode a sequence of coded symbols by guessing noise. In various embodiments, noise sequences are ordered, either during system initialization or on a periodic basis. Then, determining a codeword includes iteratively guessing a new noise sequence, removing its effect from received data symbols (e.g. by subtracting or using some other method of operational inversion), and checking whether the resulting data are a codeword using a codebook membership function. In various embodiments, soft information is used to generate a symbol mask that identifies the collection of symbols that are suspected to differ from the channel input, and only these are subject to guessing. This decoder embodies or approximates maximum likelihood (optionally with soft information) decoding for any code. In some embodiments, the decoder tests abounded number of noise sequences, abandoning the search and declaring an erasure after these sequences are exhausted.

Abstract: Described are devices, systems and techniques for implementing atomic memory objects in a multi-writer, multi-reader setting. In an embodiment, the devices, systems and techniques use maximum distance separable (MDS) codes, and may be specifically designed to optimize a total storage cost for a given fault-tolerance requirement. Also described is an embodiment to handle the case where some of the servers can return erroneous coded elements during a read operation.

Abstract: Techniques are disclosed for an adaptive and causal random linear network coding (AC-RLNC) with forward error correction (FEC) for a communication channel with delayed feedback. An example methodology implementing the techniques includes transmitting one or more coded packets in a communication channel, determining a channel behavior of the channel, and adaptively adjusting a transmission of a subsequent coded packet in the first channel based on the determined channel behavior. The communication channel may be a point-to-point communication channel between a sender and a receiver. The channel behavior may be determined based on feedback acknowledgements provided by the receiver. The subsequent coded packet may be a random linear combination of one or more information packets.

Abstract: Devices and methods described herein decode a sequence of coded symbols by guessing noise. In various embodiments, noise sequences are ordered, either during system initialization or on a periodic basis. Then, determining a codeword includes iteratively guessing a new noise sequence, removing its effect from received data symbols (e.g. by subtracting or using some other method of operational inversion), and checking whether the resulting data are a codeword using a codebook membership function. In various embodiments, soft information is used to generate a symbol mask that identifies the collection of symbols that are suspected to differ from the channel input, and only these are subject to guessing. This decoder embodies or approximates maximum likelihood (optionally with soft information) decoding for any code. In some embodiments, the decoder tests abounded number of noise sequences, abandoning the search and declaring an erasure after these sequences are exhausted.

Abstract: A network node having a receiver for receiving input packets, a local node memory where one or more parameters for coding are stored, an encoder for creating coded packets from the input packets using linear network coding, and a transmitter to transmit the coded packets. Each coefficient of the linear network coding is a parameter of the one or more parameters or a pre-determined function of the one or more parameters. A related method and a network are also presented.

Abstract: Techniques are disclosed for adaptive coding and scheduling of packets in wireless networks. The adaptive coding and scheduling can be achieved by utilizing a discrete water filling (DWF) scheme. In an example, a computer-implemented method to adaptively code and schedule packets in a wireless network may include determining number of paths between a sender and a receiver in a multipath (MP) network, determining erasure rates for each path of the paths between the sender and the receiver, and determining a multipath rate. The method may also include determining a coding bucket size based on the multipath rate and determining a multipath delay for the coding bucket size and the erasure rates. In another example, the adaptive coding and scheduling techniques can be applied to a multihop multipath (MM) network.

Abstract: Provided is a method for correcting errors in a data transmission network, comprising: transmitting a plurality of uncoded information packets across a network path; transmitting a plurality of coded packets for recovering information packets lost in transmission across said network path, the coded packets being temporally interspersed among said uncoded information packets, wherein the coded packets are encoded based on information packets transmitted prior to a previously transmitted coded packet; and determining the interspersion of the coded packets according to a packet loss rate.

Abstract: A two-layer erasure-coded fault-tolerant distributed storage system offering atomic access for read and write operations is described. In some embodiments, a class of erasure codes known as regenerating codes (e.g. minimum bandwidth regenerating codes) for storage of data in a backend layer is used to reduce a cost of backend bulk storage and helps in reducing communication cost of read operations, when a value needs to be recreated from persistent storage in the backend layer. By separating the functionality of edge layer servers and backend servers, a modular implementation for atomicity using storage-efficient erasure-codes is provided. Such a two-layer modular architecture permits protocols needed for consistency implementation to be substantially limited to the interaction between clients and an edge layer, while protocols needed to implement erasure code are substantially limited to interaction between edge and backend layers.

Abstract: Devices and methods described herein decode a sequence of coded symbols by guessing noise. In various embodiments, noise sequences are ordered, either during system initialization or on a periodic basis. Then, determining a codeword includes iteratively guessing a new noise sequence, removing its effect from received data symbols (e.g. by subtracting or using some other method of operational inversion), and checking whether the resulting data are a codeword using a codebook membership function. This process is deterministic, has bounded complexity, asymptotically achieves channel capacity as in convolutional codes, but has the decoding speed of a block code. In some embodiments, the decoder tests a bounded number of noise sequences, abandoning the search and declaring an erasure after these sequences are exhausted. Abandonment decoding nevertheless approximates maximum likelihood decoding within a tolerable bound and achieves channel capacity when the abandonment threshold is chosen appropriately.