Microsoft Technology Licensing, LLC

Amazon Technologies, Inc.

John David Dunagan

Alastair Wolman

Atul Adya

Marvin Michael Theimer

Gaurav D Ghare

Ying Xiong

Chad E Verbowski

Jiahe Helen Wang

Arne M Josefsberg

Roussi A Roussev

Mark D Scheibel

Greg Burgess

Yi-Min Wang

Malcolm Featonby

Derek Solomon Pai

John Merrill Phillips

Mohit Gupta

Mihir Sadruddin Surani

Yuxuan Liu

John Walter Heitmann

Michael R Foster

Juhan Lee

Xiaofeng Fan

Shan Lu

Arunvijay Kumar

Caitlyn Rose Shim

Gregory M Burgess

Brad Daniels

Albert Gordon Greenberg

Sumit Basu

Emre Mehmet Kiciman

Shuo Chen

Heather E Warncke

Roberto Pentz De Faria

Les Johann Lamprecht

Paramvir Bahl

Ranveer Chandra

James Robert Hamilton

Daniel Ron Simon

Sharad Agarwal

Ming Zhang

Michael B Jones

Leslie Johann Lamprecht

Jeremy Lee Dewey

Bruce Wayne Copeland

Michael J Manos

Gregory D Hartrell

Helen Jiahe Wang

Bradley Joseph Gussin

Keith A Carlson

Karthik Gopalratnam

Mansoor A Malik

Ethan John Faust

Bradford Daniels

Nicholas J A Harvey

Cary L Mitchell

Abhinav Maurya

Opher Dubrovsky

Shantanu Chandra

Kerem Bulbul

Justin J Tittelfitz

Saurabh Jain

Dejan Kostić

Azbayar Demberel

Santosh S Vempala

Erik Quanstrom

Rajesh S Batheja

James Rigby

Wenbo Bu

Alice Xiao-Zhou Zheng

Mark D Acheibel

Pawel Cholewinski

Saher Esmeir

Charles S Reis

John Carleton McCullough

David Andrew Schultz

Yaodong Liu

Wynand Jan Theunissen

Dana K Fujimoto

Dejan Kostic

A database service that receives and manages time series data into a set of generational data tables. The database service generates a set of generational data tables configured to receive a subset of time series data. Each generational data table is created with an initial set of access parameters that facilitates the receipt of incoming time series data. Thereafter, the database service creates additional generational data tables to receive the continuous received time series data and then modifies the access parameters of the previously created data table, such as for read only rights.

Generational databases for managing time series data

Techniques for predicting the availability of a resource are described. An exemplary method includes obtaining capacity data indicating an amount of capacity available in a cloud provider network to satisfy the request; generating, using a machine learning model that has been trained based at least in part on an output of an automated historical hindsight learner that is an integer linear program, an approval prediction, wherein the approval prediction indicates that the request is to be approved based on one or more launch parameters of the request and the capacity data; receiving information from a downstream component that controls the resource that the approval prediction is incorrect; and evaluating the incorrect approval prediction using a hindsight learner and predictor explainer.

Systems, methods, and apparatuses for predicting availability of a resource

Techniques for managing dynamically scalable virtualized compute instances within a provider network are described. A dynamically scalable instance has a baseline performance level and a maximum performance level associated with a computing resource, such as a processor, a memory, a network interface, etc. In response to receive a request to launch a dynamically scalable instance, one or more services select a computer system to host the requested instance from a pool of instance-hosting computer systems of a provider network. The instance-hosting computer systems include a monitoring agent that reports computing resource usage data to the one or more services. The one or more services monitor the resource usage of the instance-hosting computer systems and migrate instances so that the dynamically scalable instances can operate up to the maximum level of performance.

Adjusting dynamically scalable instance hosting based on compute resource usage

Techniques for managing compute resource usage by virtualized compute instances within a provider network are described. A computer system connected to a provider network hosts one or more compute instances. An agent associated with the computer system obtains per-compute instance resource usage information of computer system resources such as processors, memory, and network interfaces. The agent sends the usage information to a usage management service of the provider network. The usage management service generates usage limits based on the usage information from the agent and on usage targets and sends the usage limits to the computer system. The computer system limits the resource usage of the compute instance based on the usage limits.

Managing compute resource usage based on prior usage

A programmatic interface is implemented, enabling a client of a stream management service to select a data ingestion policy for a data stream. A client request selecting an at-least-once ingestion policy is received. In accordance with the at-least-once policy, a client may transmit an indication of a data record one or more times to the service until a positive acknowledgement is received. In response to receiving a plurality of transmissions indicating a particular data record, respective positive acknowledgements are sent to the client. Based on a persistence policy selected for the stream, copies of the data record are stored at one or more storage locations in response to one particular transmission of the plurality of transmissions.

Data stream ingestion and persistence techniques

A partitioning policy, comprising an indication of an initial mapping of data records of a stream to a plurality of partitions, is selected to distribute data records of a data stream among a plurality of nodes of a stream management service. Data ingestion nodes and storage nodes are configured according to the initial mapping. In response to a determination that a triggering criterion for dynamically repartitioning the data stream has been met, a modified mapping is generated, and a different set of ingestion and storage nodes are configured. For at least some time during which arriving data records are stored in accordance with the modified mapping, data records stored at the first set of storage nodes in accordance with the initial mapping are retained.

Dynamic partitioning techniques for data streams

A control node of a multi-tenant stream processing service receives a request indicating an operation to be performed on data records of a particular data stream. Based on a stream partitioning policy, the control node determines an initial number of worker nodes to be used. The control node configures a worker node to perform the operation on received records. In response to a determination that the worker node is in an unhealthy state, the control node configures a replacement worker node.

Partition-based data stream processing framework

Systems for processing requests to use virtual computing resources in communicatively isolated computing environments apply a multi-tier framework for limiting the amount of virtual computing resources that can be allocated to a user in a computing environment. A bottom tier is associated with a user account that has not been analyzed for security risks, a top tier is associated with a user account that has been validated to use a corresponding computing environment; an intermediate tier, having limits between the bottom and top tiers, is applied to computing environments where the user has not requested resources. Initially, all computing environments are on the bottom tier. In response to a valid request for resources, validated computing environments are moved to the top tier and all other computing environments are moved to the intermediate tier. Requests in an intermediate-tier computing environment trigger validity review; the computing environment can move to top tier.

Tiered usage limits across compute resource partitions

Monitoring data can be sent, received, stored, and accessed with zero setup. Customers can send monitoring data to be stored without the customers having to perform any setup ahead-of-time, such as arbitrary lists of key-value pairs. Customers can perform zero setup pivot queries using stored monitoring data. For example, users can perform pivot queries using any available stored information, such as available metric types and dimensions, without any setup needed ahead-of-time.

Flexible pivot querying of monitoring data with zero setup

A control node of a multi-tenant stream management service receives a request to initialize a data stream to be comprised of a plurality of data records. The control node determines, based on a partitioning policy, parameters to be used to configure subsystems for ingestion, storage and retrieval of the records. The control node identifies resources to be used for a node of retrieval subsystem The retrieval node is configured to implement programmatic record retrieval interfaces, including respective interfaces to implement non-sequential and sequential access patterns. The control node configures the retrieval node using the selected resources.

Managed service for acquisition, storage and consumption of large-scale data streams

Technologies, systems, and methods for ordered message delivery that avoid message races or crisscrosses between communicating nodes. For example, if Node A sends message 3 towards Node B and, shortly thereafter, Node B sends message X to Node A, Node A would like to know whether or not message X reflects Node B's state after receiving message 3. If Node B received message 3 prior to sending message X, then proper state may be maintained between the nodes. But if messages 3 and X crisscrossed, or if message 3 was never properly received by Node B, then the state between the nodes may be corrupt. Technologies, systems, and methods are provided to avoid such corruption.

Crisscross cancellation protocol

A user interface may be utilized to allow a user to control display properties of online ad content on a hosting webpage (e.g., hosting ads from one domain on a webpage from a different domain). This may be accomplished by creating a cross-domain frame in the hosting webpage, which can contain the ad content. An instantiation of a user interface may be created on the hosting webpage that has an ability to control display properties of the cross-domain frame, thereby controlling display properties of the ad content in the cross-domain frame. Further, a task manager may be utilized as part of the user interface to facilitate a host of the hosting webpage to manage ad content.

User interface for online ads

An exemplary system includes a front-end component to receive requests for resources in a data center and configured to associate each request with identifying information, to locate one or more resources for each request and to store, in a log file, the identifying information and information about the location of the one or more resources; one or more distributed computation and storage components to acquire log file information and configured to analyze log information to decide if one or more resources associated with one or more requests should be migrated to a data center in a different geographical location; and a location service component to receive decisions made by the one or more distributed computation and storage components and configured to inform the front-end component when a decision causes one or more resources to be migrated to a data center in a different geographical location to thereby allow the front-end component to re-direct future requests for the one or more migrated resources to the data center in the different geographical location. Various other devices, systems and methods are also described.

Techniques for resource location and migration across data centers

Growing community of inventors

Bellevue, WA, United States of America

John David Dunagan