Patents by Inventor Swaminathan Sivasubramanian

Swaminathan Sivasubramanian has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • ATTACHED ACCELERATOR BASED INFERENCE SERVICE
    Publication number: 20200004596
    Abstract: Implementations detailed herein include description of a computer-implemented method. In an implementation, the method at least includes receiving an application instance configuration, an application of the application instance to utilize a portion of an attached accelerator during execution of a machine learning model and the application instance configuration including: an indication of the central processing unit (CPU) capability to be used, an arithmetic precision of the machine learning model to be used, an indication of the accelerator capability to be used, a storage location of the application, and an indication of an amount of random access memory to use.
    Type: Application
    Filed: June 27, 2018
    Publication date: January 2, 2020
    Inventors: Sudipta SENGUPTA, Poorna Chand Srinivas PERUMALLA, Dominic Rajeev DIVAKARUNI, Nafea BSHARA, Leo Parker DIRAC, Bratin SAHA, Matthew James WOOD, Andrea OLGIATI, Swaminathan SIVASUBRAMANIAN
  • ATTACHED ACCELERATOR SCALING
    Publication number: 20200004597
    Abstract: Implementations detailed herein include description of a computer-implemented method.
    Type: Application
    Filed: June 27, 2018
    Publication date: January 2, 2020
    Inventors: Sudipta SENGUPTA, Poorna Chand Srinivas PERUMALLA, Dominic Rajeev DIVAKARUNI, Nafea BSHARA, Leo Parker DIRAC, Bratin SAHA, Matthew James WOOD, Andrea OLGIATI, Swaminathan SIVASUBRAMANIAN
  • Managing resources using resource expiration data
    Patent number: 10521348
    Abstract: Resource management techniques, such as cache optimization, are employed to organize resources within caches such that the most requested content (e.g., the most popular content) is more readily available. A service provider utilizes content expiration data as indicative of resource popularity. As resources are requested, the resources propagate through a cache server hierarchy associated with the service provider. More frequently requested resources are maintained at edge cache servers based on shorter expiration data that is reset with each repeated request. Less frequently requested resources are maintained at higher levels of a cache server hierarchy based on longer expiration data associated with cache servers higher on the hierarchy.
    Type: Grant
    Filed: November 2, 2018
    Date of Patent: December 31, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Bradley Eugene Marshall, Swaminathan Sivasubramanian, David R. Richardson
  • Network resource identification
    Patent number: 10511567
    Abstract: A system, method and computer-readable medium for client-side processing of resource identifiers. A client request for content is returned that includes information for generating a request for translation information and a number of embedded resource identifiers. Based on the information for generating a request for translation information, the client generates and obtains translation information and processes the embedded resource identifiers to obtain resources from a service provider.
    Type: Grant
    Filed: January 9, 2017
    Date of Patent: December 17, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: David R. Richardson, Bradley Eugene Marshall, Elmore Eugene Pope, Christopher L. Scofield, Swaminathan Sivasubramanian, John Cormie
  • Fully managed and continuously trained automatic speech recognition service
    Patent number: 10490183
    Abstract: Techniques for automated speech recognition (ASR) are described. A user can upload an audio file to a storage location. The user then provides the ASR service with a reference to the audio file. An ASR engine analyzes the audio file, using an acoustic model to divide the audio data into words, and a language model to identify the words spoken in the audio file. The acoustic model can be trained using audio sentence data, enabling the transcription service to accurately transcribe lengthy audio data. The results are punctuated and normalized, and the resulting transcript is returned to the user.
    Type: Grant
    Filed: March 15, 2018
    Date of Patent: November 26, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Ashish Singh, Deepikaa Suresh, Vasanth Philomin, Rajkumar Gulabani, Vladimir Zhukov, Swaminathan Sivasubramanian, Vikram Sathyanarayana Anbazhagan, Praveen Kumar Akarapu, Stefano Stefani
  • Managing resources and entries in tracking information in resource cache components
    Patent number: 10491534
    Abstract: A system, method, and computer readable medium for managing resources maintained in resource cache components are provided. A network storage provider storing one or more resources on behalf of a content provider obtains client computing device requests for content. The network storage provider provides resources that are received and maintained on resource cache components. The network storage provider either processes requests or provides notifications to the resource cache components to facilitate the management of resources that need to be updated or are otherwise treated as invalid.
    Type: Grant
    Filed: November 16, 2015
    Date of Patent: November 26, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: David R. Richardson, Bradley Eugene Marshall, Swaminathan Sivasubramanian
  • MANAGING RESOURCES USING RESOURCE EXPIRATION DATA
    Publication number: 20190354484
    Abstract: Resource management techniques, such as cache optimization, are employed to organize resources within caches such that the most requested content (e.g., the most popular content) is more readily available. A service provider utilizes content expiration data as indicative of resource popularity. As resources are requested, the resources propagate through a cache server hierarchy associated with the service provider. More frequently requested resources are maintained at edge cache servers based on shorter expiration data that is reset with each repeated request. Less frequently requested resources are maintained at higher levels of a cache server hierarchy based on longer expiration data associated with cache servers higher on the hierarchy.
    Type: Application
    Filed: August 2, 2019
    Publication date: November 21, 2019
    Inventors: Bradley Eugene Marshall, Swaminathan Sivasubramanian, David R. Richardson
  • MANAGING COMMUNICATIONS USING ALTERNATIVE PACKET ADDRESSING
    Publication number: 20190342212
    Abstract: Techniques are described for managing communications for a managed virtual computer network overlaid on a distinct substrate computer network, including for communications involving computing nodes of the managed virtual computer network that use an alternative addressing scheme to direct network packets and other network communications to intended destination locations by using textual network node monikers instead of numeric IP addresses to represent computing nodes at a layer 3 or “network layer” of a corresponding computer networking stack in use by the computing nodes. The techniques are provided without modifying or configuring the network devices of the substrate computer network, by using configured modules to manage and modify communications from the logical edge of the substrate network.
    Type: Application
    Filed: July 12, 2019
    Publication date: November 7, 2019
    Applicant: Amazon Technologies, Inc.
    Inventors: Swaminathan Sivasubramanian, Eric Jason Brandwine, Tate Andrew Certain, Bradley E. Marshall
  • MONITORING AND AUTOMATIC SCALING OF DATA VOLUMES
    Publication number: 20190332435
    Abstract: Aspects of a data environment, such as various capacities of data stores and instances, can be managed using a separate control environment. A monitoring component of the control environment can periodically communicate with the data environment to obtain performance information. The information is analyzed, using algorithms such as trending and extrapolation algorithms, to determine any recommended scaling of resources in the data environment. The scaling can be performed automatically, or as authorized by a customer. A workflow can be instantiated that includes tasks necessary to perform the scaling. The scaling of storage capacity can be performed without affecting the availability of the data store.
    Type: Application
    Filed: May 3, 2019
    Publication date: October 31, 2019
    Applicant: Amazon Technologies, Inc.
    Inventors: Swaminathan Sivasubramanian, Grant Alexander MacDonald McAlister, Milovan Milovanovic
  • Dynamic replication peering
    Patent number: 10461991
    Abstract: A distributed database system may comprise a master computing node and one or more replication peer nodes. Upon identifying a failure of one of the nodes, a control plane may obtain a computing node instance and configure the computing node instance to act as a log-only peer to the remaining nodes. The log-only peer may store the log tail of the master computing node and may participate in quorum decisions. Upon restoration of the failed node, the log-only peer may be returned to the pool of computing node instances.
    Type: Grant
    Filed: September 19, 2016
    Date of Patent: October 29, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Adam Douglas Morley, Timothy Andrew Rath, Swaminathan Sivasubramanian
  • Exporting dialog-driven applications to digital communication platforms
    Patent number: 10460728
    Abstract: Methods, systems, and computer-readable media for exporting dialog-driven applications to digital communication platforms are disclosed. A launch condition is received from a user. The launch condition is caused to be registered with one or more digital communication platforms. Detection of the launch condition is to cause a natural language input to be routed from at least one of the digital communication platforms to an application management service.
    Type: Grant
    Filed: June 16, 2017
    Date of Patent: October 29, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Vikram Sathyanarayana Anbazhagan, Swaminathan Sivasubramanian, Stefano Stefani, Vladimir Zhukov
  • Source detection and indexing for managed search
    Patent number: 10452675
    Abstract: A data discovery service provides an automated method of identifying and indexing data sources associated with a user's account without requiring additional configuration by the user. This enables search services to be provided across various different types of data sources a user may utilize. The data discovery service can generate a list of data sources associated with the user's account, the list may include access information for each data source. Each data source or type of data source may be associated with an index policy that identifies the types of data from a corresponding data source are to be indexed. An index may then be generated based on those policies for each data source associated with the account.
    Type: Grant
    Filed: December 30, 2015
    Date of Patent: October 22, 2019
    Assignee: Amazon Technologies, Inc.
    Inventor: Swaminathan Sivasubramanian
  • Dynamic n-dimensional cubes for hosted analytics
    Patent number: 10430438
    Abstract: An online analytical processing system may comprise an n-dimensional cube structured using slice-based partitioning in which each slice comprises one or more hierarchies of data points. A region of a hierarchy may be classified according to computational demands associated with the region. A scaling or replication mechanism may be applied to the region based on the computational demands associated with that region.
    Type: Grant
    Filed: September 23, 2014
    Date of Patent: October 1, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Santosh Kalki, Srinivasan Sundar Raghavan, Timothy Andrew Rath, Mukul Vijay Karnik, Amol Devgan, Swaminathan Sivasubramanian
  • Dynamic data set replica management
    Patent number: 10416894
    Abstract: Systems and methods dynamically manage replicas of data sets. A collection of data stores is used to redundantly store one or more replicas of one or more data sets. The replicas may be used to respond to read requests from multiple sources. Upon identification of a need to allocate storage space to a data object, space used by one or more of the replicas is allocated to the data object. Various parameters may be utilized in the selection of one or more replicas whose storage space can be allocated to the data object. The parameters may be based at least in part on characteristics of the one or more data sets relative to one another.
    Type: Grant
    Filed: October 23, 2017
    Date of Patent: September 17, 2019
    Assignee: AMAZON TECHNOLOGIES, INC.
    Inventors: Swaminathan Sivasubramanian, Bradley E. Marshall, Tate Andrew Certain, Nicholas J. Maniscalco
  • REQUEST ROUTING BASED ON CLASS
    Publication number: 20190268265
    Abstract: A system and method for management and processing of resource requests is provided. A content delivery network service provider receives a DNS query from a client computing device. The DNS query corresponds to a requested resource from the client computing device. The content delivery network service provider associates the client computing device with a cluster of other client computing devices. Based on routing information for the cluster, the content delivery network service provider routes the DNS query. The process can further include monitoring performance data associated with the delivery of the requested resource and updating the routing information for the cluster based on the performance data for use in processing subsequent requests from client computing devices in the cluster.
    Type: Application
    Filed: May 7, 2019
    Publication date: August 29, 2019
    Inventors: David R. Richardson, Swaminathan Sivasubramanian, Bradley Eugene Marshall, Christopher L. Scofield, Elmore Eugene Pope
  • Managing communications using alternative packet addressing
    Patent number: 10355991
    Abstract: Techniques are described for managing communications for a managed virtual computer network overlaid on a distinct substrate computer network, including for communications involving computing nodes of the managed virtual computer network that use an alternative addressing scheme to direct network packets and other network communications to intended destination locations by using textual network node monikers instead of numeric IP addresses to represent computing nodes at a layer 3 or “network layer” of a corresponding computer networking stack in use by the computing nodes. The techniques are provided without modifying or configuring the network devices of the substrate computer network, by using configured modules to manage and modify communications from the logical edge of the substrate network.
    Type: Grant
    Filed: September 26, 2013
    Date of Patent: July 16, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Swaminathan Sivasubramanian, Eric Jason Brandwine, Tate Andrew Certain, Bradley E. Marshall
  • Service for developing dialog-driven applications
    Patent number: 10331791
    Abstract: A natural language understanding model is trained using respective natural language example inputs corresponding to a plurality of applications. A determination is made as to whether a value of a first parameter of a first application is to be obtained using a natural language interaction. Using the natural language understanding model, at least a portion of the first application is generated.
    Type: Grant
    Filed: November 23, 2016
    Date of Patent: June 25, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: Vikram Sathyanarayana Anbazhagan, Rama Krishna Sandeep Pokkunuri, Swaminathan Sivasubramanian, Stefano Stefani, Vladimir Zhukov
  • LOCALITY BASED CONTENT DISTRIBUTION
    Publication number: 20190190998
    Abstract: A system and method for content distribution are provided. A content provider generates a network topology having one or more subnetworks made up of content sources. The content provider segments content, such as applications, into segments for distribution. The content provider then distributes the content such that for each identified subnetwork, each content segment is distributed to a content source. Subsequent, client computing device requests for content can be serviced by selecting a content source in a particular subnetwork selected for delivering the requested content to the client computing device to minimize traffic flow of the requested content through the nodes of the distribution network.
    Type: Application
    Filed: December 17, 2018
    Publication date: June 20, 2019
    Inventors: Swaminathan Sivasubramanian, David R. Richardson, Bradley Eugene Marshall
  • Request routing based on class
    Patent number: 10305797
    Abstract: A system and method for management and processing of resource requests is provided. A content delivery network service provider receives a DNS query from a client computing device. The DNS query corresponds to a requested resource from the client computing device. The content delivery network service provider associates the client computing device with a cluster of other client computing devices. Based on routing information for the cluster, the content delivery network service provider routes the DNS query. The process can further include monitoring performance data associated with the delivery of the requested resource and updating the routing information for the cluster based on the performance data for use in processing subsequent requests from client computing devices in the cluster.
    Type: Grant
    Filed: February 5, 2018
    Date of Patent: May 28, 2019
    Assignee: Amazon Technologies, Inc.
    Inventors: David R. Richardson, Swaminathan Sivasubramanian, Bradley Eugene Marshall, Christopher L. Scofield, Elmore Eugene Pope
  • FULLY MANAGED AND CONTINUOUSLY TRAINED AUTOMATIC SPEECH RECOGNITION SERVICE
    Publication number: 20190156816
    Abstract: Techniques for automated speech recognition (ASR) are described. A user can upload an audio file to a storage location. The user then provides the ASR service with a reference to the audio file. An ASR engine analyzes the audio file, using an acoustic model to divide the audio data into words, and a language model to identify the words spoken in the audio file. The acoustic model can be trained using audio sentence data, enabling the transcription service to accurately transcribe lengthy audio data. The results are punctuated and normalized, and the resulting transcript is returned to the user.
    Type: Application
    Filed: March 15, 2018
    Publication date: May 23, 2019
    Inventors: Ashish SINGH, Deepikaa SURESH, Vasanth PHILOMIN, Rajkumar GULABANI, Vladimir ZHUKOV, Swaminathan SIVASUBRAMANIAN, Vikram Sathyanarayana ANBAZHAGAN, Praveen Kumar AKARAPU, Stefano STEFANI