Sanjay Kumar Agrawal

6356890, Mar 12, 2002

Aug 1, 2000

09/629353

Sanjay Agrawal - Redmond WA
Surajit Chaudhuri - Redmond WA
Vivek R. Narasayya - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 3, 707 5

An index and materialized view selection wizard produces a fast and reasonable recommendation for a configuration of indexes, materialized views, and indexes on materialized views which are beneficial given a specified workload for a given database and database server. Candidate materialized views and indexes are obtained, and a joint enumeration of the combined materialized views and indexes is performed to obtain a recommended configuration. The configuration includes indexes, materialized views and indexes on materialized views. Candidate materialized views are obtained by first determining subsets of tables are referenced in queries in the workload and then finding interesting table subsets. Next, interesting subsets are considered on a per query basis to determine which are syntactically relevant for a query. Materialized views which are likely to be used for the workload are then generated along with a set of merged materialized views.

6356891, Mar 12, 2002

Aug 1, 2000

09/629412

Sanjay Agrawal - Redmond WA
Surajit Chaudhuri - Redmond WA
Vivek R. Narasayya - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 3, 707 5

An index and materialized view selection wizard produces a fast and reasonable recommendation for a configuration of indexes, materialized views, and indexes on materialized views which are beneficial given a specified workload for a given database and database server. Candidate materialized views and indexes are obtained, and a joint enumeration of the combined materialized views and indexes is performed to obtain a recommended configuration. The configuration includes indexes, materialized views and indexes on materialized views. Candidate materialized views are obtained by first determining subsets of tables are referenced in queries in the workload and then finding interesting table subsets. Next, interesting subsets are considered on a per query basis to determine which are syntactically relevant for a query. Materialized views which are likely to be used for the workload are then generated along with a set of merged materialized views.

6366903, Apr 2, 2002

Apr 20, 2000

09/553033

Sanjay Agrawal - Redmond WA
Surajit Chaudhuri - Redmond WA
Vivek R. Narasayya - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 4, 707 5

An index and materialized view selection wizard produces a fast and reasonable recommendation of indexes and materialized views which are beneficial given a specified workload for a given database and database server. Candidate materialized views and indexes are obtained, and a joint enumeration of the combined materialized views and indexes is performed to obtain a recommended configuration. The configuration includes both indexes and materialized views. Candidate materialized views are obtained by first determining subsets of tables are referenced in queries in the workload and then finding interesting table subsets. Next, interesting subsets are considered on a per query basis to determine which are syntactically relevant for a query. Materialized views which are likely to be used for the workload are then generated along with a set of merged materialized views.

6513029, Jan 28, 2003

Aug 1, 2000

09/629010

Sanjay Agrawal - Redmond WA
Surajit Chaudhuri - Redmond WA
Vivek R. Narasayya - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 5

An index and materialized view selection wizard produces a fast and reasonable recommendation for a configuration of indexes, materialized views, and indexes on materialized views which are beneficial given a specified workload for a given database and database server. Candidate materialized views and indexes are obtained, and a joint enumeration of the combined materialized views and indexes is performed to obtain a recommended configuration. The configuration includes indexes, materialized views and indexes on materialized views. Candidate materialized views are obtained by first determining subsets of tables are referenced in queries in the workload and then finding interesting table subsets. Next, interesting subsets are considered on a per query basis to determine which are syntactically relevant for a query. Materialized views which are likely to be used for the workload are then generated along with a set of merged materialized views.

6792414, Sep 14, 2004

Oct 19, 2001

10/036348

Surajit Chaudhuri - Redmond WA
Sanjay Agrawal - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 3

Searching by keywords and providing generalized matching capabilities on a relational database is enabled by performing preprocessing operations to construct inverted list lookup tables based on data record components at an interim level of granularity, such as column location. Prefix information is in the inverted list stored for each keyword, keyword sub-string, or stemmed version of the keyword. A keyword search is performed on the lookup tables rather than the database tables to determine database column locations of the keyword. The lookup tables is scanned to identify each prefix associated with the search term. Schema information about the database is used to link the column locations to form database subgraphs that span the keywords. Join tables are to generated based on the subgraphs consisting of columns containing the keywords. A query on the database is generated to join the tables and retrieve database rows that contain the keyword and the prefixes associated with the keyword.

6801904, Oct 5, 2004

Oct 19, 2001

10/032510

Surajit Chaudhuri - Redmond WA
Sanjay Agrawal - Redmond WA
Guatam Das - Redmond WA

Microsoft Corporation - Redmond WA

G06F 1730

707 2, 707 3

Searching by keywords on a relational database is enabled by performing preprocessing operations to construct lookup tables at an interim level of granularity, such as column location. A keyword search is performed on the lookup tables rather than the database tables to determine database column locations of the keyword. Schema information about the database is used to link the column locations to form database subgraphs that span the keywords. Join tables are to generated based on the subgraphs consisting of columns containing the keywords. A query on the database is generated to join the tables and retrieve database rows that contain the keywords. The retrieved rows are ranked in order of relevance before being output. By preprocessing a relational database to form lookup tables, and initially searching the lookup tables to obtain a targeted subset of the database upon which SQL queries can be performed to collect data records, keyword searching on relational database is made efficient.

6912547, Jun 28, 2005

Jun 26, 2002

10/180667

Surajit Chaudhuri - Redmond WA, US
Ashish Kumar Gupta - Seattle WA, US
Vivek Narasayya - Redmond WA, US
Sanjay Agrawal - Kirkland WA, US

Microsoft Corporation - Redmond WA

G06F017/30

7071041, 706917, 709224, 703 21

Relational database applications such as index selection, histogram tuning, approximate query processing, and statistics selection have recognized the importance of leveraging workloads. Often these applications are presented with large workloads, i. e. , a set of SQL DML statements, as input. A key factor affecting the scalability of such applications is the size of the workload. The invention concerns workload compression which helps improve the scalability of such applications. The exemplary embodiment is broadly applicable to a variety of workload-driven applications, while allowing for incorporation of application specific knowledge. The process is described in detail in the context of two workload-driven applications: index selection and approximate query processing.

7006440, Feb 28, 2006

Oct 26, 2001

10/047638

Sanjay K. Agrawal - San Jose CA, US
Neil N. Mammen - San Jose CA, US
Ajit Ninan - San Jose CA, US
Jason C. Fan - Mountain View CA, US

Luminous Networks, Inc. - Cupertino CA

H04L 12/26

370235

A communications network is described having a class-based queuing architecture. Shared class queues receive packet flows from different customers. In one embodiment, there are eight classes and thus eight shared queues, one for each class. A scheduler schedules the output of packets by the various queues based on priority. Each customer (or other aggregate of packet flows) is allocated a certain space in a class queue based on the customers' Service Level Agreement (SLA) with the service provider. A queue input circuit detects bits in the packet header identifying the customer (or other criteria) and makes selections to drop or pass packets destined for a shared queue based on the customers' (or other aggregates') allocated space in the queue. In another embodiment, the relative positions of the nodes in the network are taken into account by each node when dropping packets forwarded by other nodes by detecting a node label (or other ID code) so that packets from the various nodes are dropped in a more fair way when there is congestion in the network, irrespective of the “passing ” node's position relative to the other nodes.