Micro Focus Lcc

Yuanzhe Bei

Alexander Kalinin

Thao Nguyen Pham

Michael Leuchtenburg

Ning Deng

Charles Edward Bear

Hoa Binh Nga Tran

Niccolo Meneghetti

Nga Tran

Olga Papaemmanouil

Chi Zhang

Manasee Moghey

Devosh Mathivanan

Zhibo Peng

Zhihao Zheng

A distributed database system maintains a database including a data shard for which a primary computing node is responsible. The primary computing node identifies a data storage plan for the data shard. The plan identifies a file subset of data storage files of the shard to be merged into a larger data storage file, and a node subset of computing nodes of the system that subscribe to the data shard. The primary node identifies which computing nodes of the node subset each have sufficient computing resources to execute the plan, as candidate computing nodes. The primary node identifies which files of the file subset each candidate computing node locally caches. The primary node selects one candidate computing node to execute the plan, based on the files of the file subset that each candidate computing node locally caches. The primary node causes the selected candidate computing node to execute the plan.

Selecting subscribing computing node to execute data storage plan for data shard

In multi-threaded or multi-processor computing systems, a deadlock may occur when two or more processes or threads are unable to proceed because they are each waiting for a resource that the other holds. As a result, progress is halted because conflicting entities are stuck in a circular dependency, and none can release the resources they hold to let the others continue. Systems and methods are provided wherein a resource reservation is carried out in two steps. The first step causes query nodes to add an identifier to a queue and, upon a request and the identifier being in a first position, a non-sharable resource is reserved. As a result, non-sharable resources are reserved in order and when needed, thereby preventing deadlocks.

Deadlock prevention utilizing distributed resource reservations

Current physical resources utilization of a computing system as a whole is monitored. The number of queries concurrently being executed against a database by a database management system (DBMS) running on a computing system is monitored. A query plan for a received query to be executed against the database is generated. The query plan includes operators; the generation of the query plan includes generation of query-based statistics for the received query on a per-operator basis without consideration of the queries concurrently being executed. An estimated execution time of the received query is dynamically predicted using a machine-learning model based on the query-based statistics generated for the received query on the per-operator basis, the current physical resources utilization of the computing system, and the number of queries concurrently being executed. The received query is executed against the database based on the dynamically predicted estimated execution time for the received query.

Machine learning model-based dynamic prediction of estimated query execution time taking into account other, concurrently executing queries

A query plan includes steps to implement a query and debug steps interleaved among the steps. An execution engine of a database system executes each step of the query plan to realize the query. The execution engine executes each debug step of the query plan to generate debugging information for the step preceding the debug step within the query plan. The debugging information is queryable.

Generating debugging information for query plan steps

In some examples, asymmetric quorum protocol based distributed transaction database consistency control may include determining, by a set of primary participants of a cluster, whether a cluster-wide quorum has been reached to commit a transaction associated with a distributed transactional database. Further, based on the determination by the set of primary participants that the cluster-wide quorum has been reached to commit the transaction, instructions to a set of secondary participants of the cluster may be sent to commit the transaction. The set of secondary participants may be configured to comply with a determination by the set of primary participants to commit or to not-commit the transaction.

Asymmetric quorum protocol based distributed transaction database consistency control

In some examples, candidate projection enumeration based query response generation may include obtaining a query, and identifying, based on the query, a plurality of candidate columns from at least one table to respond to the query. Using a tree-based model, the candidate columns may be analyzed to generate a candidate set of projections to respond to the query. Each candidate projection of the candidate set of projections may include a sort order associated with at least one candidate column of the candidate columns, a segmentation key value associated with at least one candidate column of the candidate columns, and a replication value associated with the candidate projection. Based on the candidate set of projections, a response to the query may be generated.

Candidate projection enumeration based query response generation

According to an example, a system may receive a refresh column command for a column-oriented database. The refresh column command may identify a table in the column-oriented database and a denormalized column in the table. The system may determine column definitions for the column, generate a query to populate values from a data source associated with the column into the column based on the column definitions, and may execute the query to replace entire set of existing column values with the data source values.

Database column refresh via replacement

A flattened table (FT) of a database of a database management system (DBMS) is defined. The FT logically materializes a number of columns over a number of partitions. The columns include normalized columns, denormalized columns from a plurality of source tables of the database, as well as an aggregate column defining an aggregation of a selected normalized column over a selected denormalized column. A live-aggregate projection (LAP) is defined on the FT of the database. The LAP corresponds to the aggregate column and physically materializes the aggregation of the selected denormalized column over the selected denormalized column, as defined by the aggregate column. The FT is refreshed on-demand, on a per-column, per-partition basis. Responsive to the FT being refreshed on-demand, the LAP is automatically refreshed.

Growing community of inventors

Santa Clara, CA, United States of America

Yuanzhe Bei