Box Limited

Hrl Laboratories, LLC

Kirill Minkovich

Karim El Defrawy

Joshua W Baron

Eric P Tressler

Narayan Srinivasa

Aleksey Nogin

Pranava Kumar Adduri

Xiaodong Wang

Heiko Hoffmann

Gavin D Holland

Corey M Thibeault

Mark Storer

Advait Kumar Mishra

Edward Chan

Mohammed Alhalalsheh

Mindy Yang

Yougkwan Cho

Girish Thattil

Vova Galchenko

Luis Miguel Hernanz Iglesias

Susmitha Kanakamedala

Embodiments of the disclosure provide systems and methods for reducing the capacity used to provide High Availability (HA) and Disaster Recovery (DR) in a distributed computing environment. According to one embodiment, dynamic recovery of a cloud-based resource can comprise setting a current latency value to an initial latency value and handling received requests with the current latency value. Current resource utilization can be detected while requests are being processed and a determination can be made as to whether the detected current resource utilization exceeds a predetermined threshold amount of resource utilization. In response to determining the detected current resource utilization does not exceed the threshold, the current latency amount can be maintained at the initial latency value. In response to determining the detected current resource utilization exceeds the threshold, the current latency value can be adjusted and injected into handling of received client requests.

Methods and systems for dynamic recovery with adaptive latency injection

Embodiments are directed to association of storage objects in a multi-tenant, cloud-based storage environment. More specifically, a first object can be made available in the file system of the environment and can be accessed, for example, by its file or object name. At some point thereafter, an auxiliary object can be provided that is related to the first object. For example, this auxiliary object can comprise metadata or other content associated with or related to the first object. According to one embodiment, this auxiliary object can be assigned an identifier that not only uniquely identifies the auxiliary object but also identifies, within the identifier itself, the relationship to the first object. This relational naming can be used to track the relationship between objects, e.g., to update the auxiliary object based on changes to the first object, but without a separate mapping of the objects.

Methods and systems for storing and associating objects in a storage system

Embodiments of the disclosure provide systems and methods for providing random access to segmented and encrypted or compressed data stored in a repository. Retrieving at least a portion of a file stored in a repository can comprise storing a plurality of files in the repository. A request to retrieve at least a portion of one of the plurality of files can be received and object metadata for the requested one of the plurality of files can be obtained. A determination can be made based on the obtained metadata as to whether the requested one of the plurality of files is a multipart file. In response to determining the requested one of the plurality of files is not a multipart file, a single file retrieval process performing and in response to determining the requested one or the plurality of files is a multipart file, a multipart retrieval process can be performed.

Random access for chunked and compressed files

Embodiments of the disclosure provide systems and methods for identifying hierarchically related files stored in a database. According to one embodiment, this can be accomplished using a pipeline of worker applications which each use a specific table stored in the databaseand a sequence or series of queues. Generally speaking, each worker application can read a message identifying tasks to be performed from one queue in the sequence of queues, perform those tasks using the appropriate table from the database to identify a set of files, and write one or more messages to a subsequent queue in the sequence or series of queues to pass results and/or further tasks to be performed to a subsequent worker application in the pipeline of worker applications.

Methods and systems for identifying and retrieving hierarchically related files

Disclosed is an approach to incorporate geographical access control features for a cloud-based storage platform. This allows, for example, enterprise administrators to define geographical areas (geofences) with arbitrary precision within which content access can be denied for items of data.

Geofencing of data in a cloud-based environment

Disclosed is an approach to incorporate geographical access control features for a cloud-based storage platform. This allows, for example, enterprise administrators to define geographical areas (geofences) with arbitrary precision within which content access can be applied to items of data.

Described is a system for a cloud control operations plane. In operation, a job is broadcast to a plurality of physical hosts, one or more of the physical hosts having a control operations plane (COP) node and a service node associated with the COP node. The COP nodes jointly create a private job assignment. A set of job assignments is redundantly distributed to individual COP nodes pursuant to the private job assignments, such that each individual COP node is only aware of its own assignment and corresponding job. The service nodes then each complete a task associated with the job and generate an output. When a set of service nodes performing a redundant job complete their task, the corresponding COP nodes jointly perform a private result checking protocol to generate a final output. The final output is then sent to the user.

System and method for cloud control operations plane based on proactive security algorithms

Described is a protocol for multi-dimensional secure pattern matching. The protocol is to be evaluated between two parties, P(or Client) and P(or Server). Pholds a multi-dimensional pattern, p, and Pholds a multi-dimensional text T (where both p and T have the same number of dimensions, but where p may be of smaller length in each dimension compared to T). Pand Pwould then engage in a protocol that allows Pto find out whether p is present in T or not. The security and privacy requirements are that Pdoes not learn any information about the pattern p nor the result of the matching. Pshould also not learn any information about T other than whether p is present in it or not. Upon implementation of the protocol, p matches T if there exists an m&#xd7; . . . &#xd7;m sub-hypermatrix (or sub-array) of T that equals p.

Secure multi-dimensional pattern matching for secure search and recognition

A neural network portion comprising N pre-synaptic neurons capable each of firing an action potential, wherein the number N can be encoded in a word of n bits;

Firing rate independent spike message passing in large scale neural network modeling

A neuromorphic compiler includes a placement module to provide analytic placement of neurons in a neural network description. The analytic placement is to produce placed neurons. The neuromorphic compiler further includes a smoothing module to perform diffusion-based smoothing of the placed neurons; a legalization module to adjust locations of the placed neurons to correspond to legal locations of neuromorphic neurons within a neural fabric; and a simulated annealing module to refine locations of the placed neurons within the neural fabric using simulated annealing following location adjustment by the legalization module. The neural fabric is to implement a synaptic time-multiplexed (STM) neuromorphic network.

Neuromorphic compiler

Described is a system for allowing sets of processors to engage in a secure pattern matching protocol. An input pattern is received from a first set of processors, while a text is received from a second set of processors. A matrix is constructed based on values computed for each character determined by each character's position in the pattern. The first set of processors sends an encrypted matrix to the second set of processors. The second set of processors processes each character in the text and creates a set of vectors. A final activation vector is created based on processing the set of vectors and an encrypted activation vector. The second set of processors sends the final activation vector to the first set of processors. The second set of processors decrypts the final activation vector. The system provides to the first set of processors where the pattern matches the text.

Growing community of inventors

Campbell, CA, United States of America

Kirill Minkovich