Amazon Technologies, Inc.

Andrew Evenson

Ayan Kumar

Samuel Leonard Moniz

Keian Christopher

Krishnan Ananthanarayanan Kolazhi

Onkar Bhaskar Walavalkar

Khawar Ali

Ruogu Zhang

Yue Yu

Kyle Bradley Peterson

Kevin Michael Beranek

Christopher Keakini Kaulia

Catalin Mihai Constantin

Aman Gupta

Christopher Chandler

Abhishek Arora

Xuetao Fan

Robert Mcgregor Calhoun, Jr

Addison Phillips

Cheng Shao

David Kerch Weber

Jiri Mares

Masud Ahmed Khan

Jeremy David McCarthy

Himesh Kumar

Muralidharan Subramani

Kevin Kwon

Michael Sutton

Viktor Iourievich Tarassov

Zhongwei Yao

Ruoshi Li

Dmitri Gluzman

Hari Chandana Kanchanapally

Methods, systems, and computer-readable media for delivery of log records to stateless clients are disclosed. A record delivery system receives, from a client, a first request to read from a persistent log comprising an ordered sequence of records. The first request is associated with a receiver session. The system sends a first set of records to the client and stores a data structure indicating that the first set of records was sent to the client in the receiver session. The system receives, from the client, a second request to read from the persistent log in the receiver session. Based at least in part on the data structure, the system determines a second set of one or more records in the persistent log. The system sends the second set of records to the client.

Delivery of log records to stateless clients

An index handler determines, with respect to a key to be inserted into an index, that a candidate destination node of the index meets a split criterion. The index handler generates and embeds a deferred split descriptor comprising an identifier of a new node within the destination node. Before an insert-completed indication is provided, the destination node is written to a back-end data store without acquiring a lock and without writing out the new node to the back-end data store. During the traversal of the index, the index handler identifies another deferred split descriptor indicating a second new node. After providing the indication that the key was successfully inserted, the index handler writes the second new node to the back-end data store.

Immediately-consistent lock-free indexing for distributed applications

Techniques are described for generating transaction data for a transaction, the transaction data generated with varying fidelity levels based on various constraints. In response to a transaction request from a client, a contract engine may perform service call(s) that instruct network service(s) to generate transaction data. The service call(s) may include constraint(s) on the generation of the transaction data, such as a time limit or a requested fidelity. The network service(s) may generate the transaction data at a particular fidelity based on the time limit, requested fidelity, or other criteria such as current load on the network service(s). The transaction data may be communicated to the contract engine with a fidelity indicator. Lower fidelity transaction data may be sent to the client to enable a low latency user experience. The lower fidelity transaction data may subsequently be replaced with higher fidelity transaction data generated with higher latency.

Method, computer-readable media, and system for transaction data generation with varying fidelity

Systems, devices, and methods are provided for fingerprinting requests, such as transaction records. A transaction record or other suitable request may be parsed to identify a parameter values for a set of fields relevant to fingerprinting. A transaction record representation may be generated based on transaction input data, transaction output data, intermediate data, or combinations thereof. A fingerprint may be generated from the transaction record implementation. Fingerprints may be used to identify various test cases that can be used for regression testing.

Test case generation using fingerprinting

Methods and systems for implementing a highly available distributed queue using replicated messages are disclosed. An enqueue request is received from a client at a particular queue host of a plurality of queue hosts. The enqueue request comprises a message and a replica count greater than one. One or more copies of a replication request are sent from the particular queue host to one or more additional queue hosts. The replication request comprises the message. The quantity of copies of the replication request is determined based at least in part on the replica count. An initial replica of the message is enqueued at the particular queue host. One or more additional replicas of the message are enqueued at the one or more additional queue hosts. A quantity of the one or more additional replicas is determined based at least in part on the replica count.

Highly available distributed queue using replicated messages

Methods, systems, and computer-readable media for highly available storage using independent data stores are disclosed. A request to create a data object using an atomic operation is received. A selected data store is determined of a plurality of data stores comprising a first data store and a second data store. The data stores are mutually independent. An identifier assigned to the data object indicates the selected data store. An event associated with creation of the data object is generated. The event is associated with the identifier. The selected data store stores the data object and the event such that the event is consistent with the data object.

Highly available storage using independent data stores

Methods, systems, and computer-readable media for safe deployment of configurations to server fleets are disclosed. A host processes requests to a service according to a control versions of a first configuration and a second configuration. After receiving an experimental version of the first configuration, a host processes some requests to the service according to the control versions of the first configuration and the second configuration, and other requests according to the experimental version of the first configuration and the control version of the second configuration. The experimental version of the first configuration is approved or rejected based at least in part on performance of the service. The host then processes some requests to the service according to the experimental version of the first configuration and the control version of the second configuration, and other requests according to the experimental versions of the first configuration and the second configuration.

Safe deployment of configurations to server fleets

A distributed hash table (DHT) based logging service utilizes a DHT to store an event log or another type of data. For example, applications can utilize the logging service to store data that logs aspects of their operation. A key/value storage service can provide access to the DHT. The data stored in the DHT can be organized into a 'skipchain' data structure. The data structure includes nodes storing the data that are distributed across partitions of the DHT. Nodes in the data structure can identify 'redundancy nodes' and &#x201c;skip nodes.&#x201d; When a request is received to add new data to the skipchain data structure, a new node is stored in the DHT that includes the new data. A head node for the data structure is updated in the DHT. The head node can also be utilized to recover from a failure to write an internal node to the data structure.

Distributed hash table based logging service

Techniques for determining the relationships between a network of services include determining encoding of text fields received and transmitted by the services. Correspondence between the bytes representing a text string in a first data field and encoding data that associates byte characteristics with encoding schemes may be used to determine the encoding scheme associated with the first data field. Other data fields that were transmitted or received close-in-time to the first data field and have a value within a threshold similarity of that of the first data field may be determined to be associated with the first data field. Subsequent to a modification to the network of services, if the encoding of one or more data fields changes or if the associations between data fields change, the particular services that were affected by the modification may be determined based on the differences in encoding schemes and data field associations.

System for differentiating encoding of text fields between networked services

Methods, systems, and computer-readable media for chain logging using key-value data storage are disclosed. A chain is generated that represents a plurality of events in chronological order. The chain comprises a head and one or more nodes. The head and the one or more nodes are stored in a persistent data store as a plurality of key-value pairs. For a request to add data indicative of a new event to the chain, the head is updated in the data store to include the data indicative of the new event.

Chain logging using key-value data storage

Methods, systems, and computer-readable media for globally ordered event stream logging are disclosed. A first host of a plurality of hosts generates an additional chain of a stream comprising a plurality of chains and representing a globally ordered sequence of events. The first host sends information describing the additional chain to a second host. The second host receives a request to store a data object representing an event. The request is associated with a stream identifier of the stream. The second host stores, in the additional chain, the data object, a timestamp associated with the data object, and an identifier of the second host. The data object in the additional chain is associated with a position in the globally ordered sequence across the plurality of chains.

Globally ordered event stream logging

A distributed hash table (DHT) based logging service utilizes a DHT to store an event log or another type of data. For example, applications can utilize the logging service to store data that logs aspects of their operation. A key/value storage service can provide access to the DHT. The data stored in the DHT can be organized into a 'skipchain' data structure. The data structure includes nodes storing the data that are distributed across partitions of the DHT. Nodes in the data structure can identify 'redundancy nodes' and 'skip nodes.' When a request is received to add new data to the skipchain data structure, a new node is stored in the DHT that includes the new data. A head node for the data structure is updated in the DHT. The head node can also be utilized to recover from a failure to write an internal node to the data structure.

Growing community of inventors

Seattle, WA, United States of America

Andrew Evenson