At&t Intellectual Property I, LP

Marios Hadjieleftheriou

Divesh Srivastava

Theodore Donald Johnson

Vladislav Shkapenyuk

Nick Koudas

Graham Cormode

Ke Yi

Xiaohui Yu

George Kollios

Feifei Li

Amit Chandel

In embodiments of the disclosed technology, indexes, such as inverted indexes, are updated only as necessary to guarantee answer precision within predefined thresholds which are determined with little cost in comparison to the updates of the indexes themselves. With the present technology, a batch of daily updates can be processed in a matter of minutes, rather than a few hours for rebuilding an index, and a query may be answered with assurances that the results are accurate or within a threshold of accuracy.

Incremental maintenance of inverted indexes for approximate string matching

Concepts and technologies are disclosed herein for managing a distributed database. A data management application can obtain a query. The data management application can analyze the query to determine a number of data structures relevant to the query. The data management application also can analyze data stores storing the data structures and move or assign data structures to other data stores within a distributed database. The movement of the data structures within the distributed database can be based upon greedy algorithms for moving data and/or executing queries.

Managing a distributed database

Concepts and technologies are disclosed herein for generating and using temporal metadata partitions. Metadata can be stored in temporal metadata partitions based upon a time range included in the metadata. Furthermore, metadata can be stored in multiple temporal metadata partitions to which the metadata is relevant. As such, metadata can be stored in manner that allows event data to be understood in the context of temporally accurate and/or relevant metadata. Functionality for executing queries of event data and providing results in view of metadata, as well as the merging of multiple temporal metadata partitions also are disclosed.

Generating and using temporal metadata partitions

Concepts and technologies are disclosed herein for generating and using temporal data partition revisions. In some embodiments, a server computer can execute a data management application. The server computer can obtain a data partition associated with a data stream. The data partition can include a data point and a timestamp. The server computer can analyze the timestamp to determine if the data partition corresponds to an update of a base data record. If the server computer determines that the data corresponds to an update, the server computer can compute a revision comprising a delta and a timestamp and store the revision.

Generating and using temporal data partition revisions

The invention relates to a system and/or methodology for selectivity estimation of set similarity queries. More specifically, the invention relates to a selectivity estimation technique employing hashed sampling. The invention providing for samples constructed a priori that can efficiently and quickly provide accurate estimates for arbitrary queries, and can be updated efficiently as well.

Selectivity estimation of set similarity selection queries

Embodiments disclosed herein are directed to verifying query results of an untrusted server. A data owner outsources a data stream to the untrusted server, which is configured to respond to a query from a client with the query result, which is returned to the client. The data owner can maintain a vector associated with query results returned by the server and can generate a verification synopsis using the vector and a seed. The verification synopsis includes a polynomial, where coefficients of the polynomial are determined based on the seed. The data owner outputs the verification synopsis and the seed to a client for verification of the query results.

Verification of outsourced data streams

The similarity between a query set comprising query set tokens and a database set comprising database set tokens is determined by a similarity score. The database sets belong to a data collection set, which contains all database sets from which information may be retrieved. If the similarity score is greater than or equal to a user-defined threshold, the database set has information relevant to the query set. The similarity score is calculated with an inverse document frequency method (IDF) similarity measure independent of term frequency. The document frequency is based at least in part on the number of database sets in the data collection set and the number of database sets which contain at least one query set token. The length of the query set and the length of the database set are normalized.

Growing community of inventors

Morristown, NJ, United States of America

Marios Hadjieleftheriou