Oracle International Corporation

Hewlett-packard Development Company, L.p.

Shasank Kisan Chavan

Allison Holloway

Jesse Kamp

Amit Ganesh

Vineet Marwah

Tirthankar Lahiri

Aurosish Mishra

Michael J Gleeson

Anindya Chandra Patthak

Teck Hua Lee

Weiwei Gong

Dennis Lui

Laurent Daynes

Vlad Ioan Haprian

Jean-Pierre Lozi

Marco Arnaboldi

Zhen Hua Liu

Sungpack Hong

Garret Frederick Swart

Ajit Mylavarapu

Hariharan Lakshmanan

Phumpong Watanaprakornkul

Sheldon A K Lewis

Niloy Mukherjee

Manosiz Bhattacharyya

Roger MacNicol

Prashant Gaharwar

Lawrence J Ellison

William Martinez Cortes

Juan R Loaiza

Vivekanandhan Raja

Vasileios Trigonakis

Sanket Hase

Hugo Kapp

Krishna Kunchithapadam

Xia Hua

Dina Thomas

Chinmayi Krishnappa

Victor Chen

Roger Dermot Macnicol

Ciprian Baetu

Meichun Hsu

Yunrui Li

Boris Klots

Siva Ravada

Vikas Arora

Ying Hu

Vasudha Krishnaswamy

Hideaki Kimura

Sukhada Pendse

Kartik Kulkarni

Viral Shah

James Kearney

Maria Colgan

Kamaljit Shergill

Jiaqi Yan

Andrew Witkowski

Robert T Golla

Mohamed Zait

Jeffrey S Brooks

Kothanda Umamageswaran

Douglas James McMahon

Krishnan Meiyyappan

Hui Joe Chang

Sujatha Muthulingam

Sunil Parrokkaren Chakkappen

Ramesh Kumar

Paul Del Vigna, Jr

Adrian Tsz Him Ng

David L Weaver

You Jung Kim

Adam Kociubes

Ekrem S C Soylemez

Nirav Vyas

Vinita Subramanian

Eugene Karichkin

Ravi Palakodety

Dhruvil Shah

Danica Porobic

Shubhro Jyoti Roy

Yu Chieh Fu

Ken Kumar

Yiran Qin

Jim Kearney

Albert Danysh

Prateek Agrawal

Andrew Ewoldt

Can Tuzla

Andrea Segalini

Sheldon Lewis

Mahendra Maiti

Sooyeon Jo

Vicente Hernandez Jimenez

Gary Smith

Vicente Hernandez

Saraswathy Narayan

Lawrence C Jenkins

Sangho Lee

Roger D Macnicol

Douglas J Mcmahon

Techniques are provided for optimizing storage of database records in segments using sub-segments. A base segment is a container used for storing records that belong to a database object. A database management system receives a request to load, into the database object, a first set of records that are in a first state. In response to receiving the request, the system generates a new sub-segment, which is a container that is separate from the base segment. The system stores the first set of records, in their first state, within the sub-segment. The system then monitors one or more characteristics of the database system. In response to the one or more characteristics satisfying criteria, the system performs a migration of one or more records of the first set of records from the sub-segment to the base segment. During the migration, the system converts the one or more records from the first state to a second state and stores the one or more records, in their second state, in the base segment.

Automatic, online and space efficient clustering and compression of very large database objects

Techniques are provided for determining an optimal configuration for an in-memory store based on both benefits and overhead that would result from having database elements populated in the in-memory store. The techniques include determining an overhead-adjusted benefit score for each element based, at least in part, on (a) a scan-benefit value, (b) a scan-overhead value, and (c) a DML-overhead value. Based on the plurality of overhead-adjusted benefit scores, the database determines an optimal configuration of the in-memory store, and then evicts in-memory copies of elements and/or loads in-memory copies of elements based on the optimal configuration.

Optimize in-memory column store for mixed workloads

Techniques for accelerating query execution by optimizing data transfer between storage nodes and database nodes are provided. In one technique, a compute node receives a database statement and transmits a set of one or more selection criteria associated with the database statement to a storage node. Based on the database statement, the storage node retrieves a set of data blocks from storage. Each data block comprises multiple rows of an index-organized table (IOT), each row comprising a key section and a non-key section. The storage node applies the set of selection criteria to a data block, resulting in a modified data block. The storage node generates a modified header data for the modified data block and transmits the modified data block to the compute node.

Accelerating query execution by optimizing data transfer between storage nodes and database nodes

Techniques are provided for optimizing workload performance by automatically discovering and implementing performance optimizations for in-memory units (IMUs). A system maintains a set of IMUs for processing database operations in a database. The system obtains a database workload information for the database system and filters the database workload information to identify database operations in the database workload information that may benefit from performance optimizations. The system analyzes the database operations to identify a set of performance optimizations and ranks the performance optimizations based on their potential benefit. The system selects a subset of the performance optimizations, based on their ranking, and generates new versions of IMUs that reflect the performance optimizations. The system performs verification tests on the new versions of IMUs and analyzes the tests to determine whether the new versions of IMUs yield expected performance benefits. The system then categorizes the new set of IMUs into a first set of IMUs to be retained and a second set of IMUs to be discarded. The system then makes the first set of IMUs available to the current workload and discards the second set of IMUs.

Optimize workload performance by automatically discovering and implementing in-memory performance features

Techniques related to cache storage formats are disclosed. In some embodiments, a set of values is stored in a cache as a set of first representations and a set of second representations. For example, the set of first representations may be a set of hardware-level representations, and the set of second representations may be a set of non-hardware-level representations. Responsive to receiving a query to be executed over the set of values, a determination is made as to whether or not it would be more efficient to execute the query over the set of first representations than to execute the query over the set of second representations. If the determination indicates that it would be more efficient to execute the query over the set of first representations than to execute the query over the set of second representations, the query is executed over the set of first representations.

Storage formats for in-memory caches

The present invention relates to optimized access of a database. Herein are techniques to accelerate execution of any combination of ad hoc query, heterogenous hardware, and fluctuating workload. In an embodiment, a computer receives a data access request for data tuples and compiles the data access request into relational operators. A particular implementation of a particular relational operator is dynamically selected from multiple interchangeable implementations. Each interchangeable implementation contains respective physical operators. A particular hardware operator for a particular physical operator is selected from multiple interchangeable hardware operators that include: a first hardware operator that executes on first processing hardware, and a second hardware operator that executes on second processing hardware that is functionally different from the first processing hardware. A response to the data access request is generated based on: the data tuples, the particular implementation of the particular relational operator, and the particular hardware operator.

Techniques of heterogeneous hardware execution for SQL analytic queries for high volume data processing

A shared-nothing database system is provided in which parallelism and workload balancing are increased by assigning the rows of each table to 'slices', and storing multiple copies (“duplicas”) of each slice across the persistent storage of multiple nodes of the shared-nothing database system. When the data for a table is distributed among the nodes of a shared-nothing system in this manner, requests to read data from a particular row of the table may be handled by any node that stores a duplica of the slice to which the row is assigned. For each slice, a single duplica of the slice is designated as the “primary duplica”. All DML operations (e.g. inserts, deletes, updates, etc.) that target a particular row of the table are performed by the node that has the primary duplica of the slice to which the particular row is assigned. The changes made by the DML operations are then propagated from the primary duplica to the other duplicas (“secondary duplicas”) of the same slice.

System and method for an ultra highly available, high performance, persistent memory optimized, scale-out database

Techniques for processing top-K queries are provided. In one technique, a database statement is received that requests top-K results related to a database object and that indicates two columns thereof: a first column by which to partition a result set and a second column by which to order the result set. A buffer is generated. For each of multiple rows in the database object: a first key value that associated with a first value in the first column of said each row is identified; a second key value that associated with a second value in the second column of said each entry is identified; a slot in the buffer is identified based on the first key value and the second key value; and the slot in the buffer may be updated based on the second key value. A response to the database statement is generated based on the buffer.

Processing top-K queries on data in relational database systems

Herein are techniques for dynamic aggregation of results of a database request, including concurrent grouping of result items in memory based on quasi-dense keys. Each of many computational threads concurrently performs as follows. A hash code is calculated that represents a particular natural grouping key (NGK) for an aggregate result of a database request. Based on the hash code, the thread detects that a set of distinct NGKs that are already stored in the aggregate result does not contain the particular NGK. A distinct dense grouping key for the particular NGK is statefully generated. The dense grouping key is bound to the particular NGK. Based on said binding, the particular NGK is added to the set of distinct NGKs in the aggregate result.

Cache conscious techniques for generation of quasi-dense grouping codes of compressed columnar data in relational database systems

The present invention relates to join acceleration. In an embodiment, a computer receives a request for a relational join of build data rows with probe data rows. Based on the request for the relational join, a particular kind of data map from many kinds of data map that can implement the relational join is dynamically selected. Based on the build data rows, an instance of the particular kind of data map is populated. A response is sent for the request for the relational join that is based on the probe data rows and the instance of the particular kind of data map.

Hardware friendly key value table design to support relational hash join

Described are improved systems, computer program products, and methods for adaptively provisioning an ordered sequence in a clustered database environment. The approach includes identifying a cached list of sequence numbers. A request for one or more sequence numbers in a database environment may be received. A determination may be made to decide whether the request pertains to an ordered sequence. The one or more sequence numbers may be adaptively provisioned to fulfill the request based at least in part upon an independent cache implementation or upon a shared cache implementation.

Methods, systems, and computer program products for adaptively provisioning a sequence in a clustered database environment

The present invention relates to hash join acceleration. In an embodiment, a computer receives a request to join build data rows with probe data rows. Each data row is associated with a respective join key value. From multiple buckets of a hash table, a bucket is selected that contains entries of multiple respective join key values in respective locations within the bucket. Whether or not the bucket contains an entry of the join key value of a build data row is detected. While building, multiple locations within the bucket are concurrently inspected to find an empty location in which to store the entry of the join key value of the build data row. The bucket is concurrently probed to detect whether or not the entries in the bucket include an entry of a join key value of a probe data row. For the join request, a response is sent that is based on the concurrent probing of the bucket.

Accelerated building and probing of hash tables using symmetric vector processing

Techniques herein use in-memory column vectors to process data that is external to a database management system (DBMS) and logically join the external data with data that is native to the DBMS. In an embodiment, a computer maintains a data dictionary for native data that is durably stored in an DBMS and external data that is not durably stored in the DBMS. From a client through a connection to the DBMS, the computer receives a query. The computer loads the external data into an in-memory column vector that resides in random access memory of the DBMS. Based on the query and the data dictionary, the DBMS executes a data join of the in-memory column vector with the native data. To the client through said connection, the computer returns results of the query based on the data join.

Performing in-memory columnar analytic queries on externally resident data

Herein are techniques that concurrently populate entries in a compressed sparse row (CSR) encoding, of a type of edge of a heterogenous graph. In an embodiment, a computer obtains a mapping of a relational schema to a graph data model. The relational schema defines vertex tables that correspond to vertex types in the graph data model, and edge tables that correspond to edge types in the graph data model. Each edge type is associated with a source vertex type and a target vertex type. For each vertex type, a sequence of persistent identifiers of vertices is obtained. Based on the mapping and for a CSR representation of each edge type, a source array is populated that, for a same vertex ordering as the sequence of persistent identifiers for the source vertex type, is based on counts of edges of the edge type that originate from vertices of the source vertex type. For the CSR, the computer populates, in parallel and based on said mapping, a destination array that contains canonical offsets as sequence positions within the sequence of persistent identifiers of the vertices.

Parallel and efficient technique for building and maintaining a main memory CSR based graph index in a RDBMS

Growing community of inventors

Menlo Park, CA, United States of America

Shasank Kisan Chavan