Page
%P
-
Article
Joins on high-bandwidth memory: a new level in the memory hierarchy
High-bandwidth memory (HBM) gives an additional opportunity for hardware performance benefits. The high available bandwidth compared to regular DRAM allows execution of many threads in parallel, avoiding memor...
-
Chapter
A Survey of B-Tree Locking Techniques
B-trees have been ubiquitous in database management systems for several decades, and they are used in other storage systems as well. Their basic structure and basic operations are well and widely understood in...
-
Chapter
On Transactional Concurrency Control
This volume contains a number of chapters on transactional database concurrency control. A two-sentence summary of the volume’s entire sequence of chapters is this: traditional locking techniques can be improv...
-
Chapter
A Problem in Two-Phase Commit
A common optimization in distributed databases lets read-only sites participate only in the 1st phase of a 2-phase commit. Upon receiving the request to vote from the commit coordinator, read-only sites releas...
-
Chapter
The End of Optimistic Concurrency Control
This chapter argues that all database systems and key-value stores should remove any implementations of optimistic concurrency control, including those based on timestamp validation. For less contention and le...
-
Chapter
Concurrent Queries and Updates in Summary Views and Their Indexes
Materialized views have become a standard technique in decision support databases and for a variety of monitoring purposes.1 In order to avoid inconsistencies and thus unpredictable query results, materialized vi...
-
Chapter
Orthogonal Key-Value Locking
B-trees have been ubiquitous for decades in databases, file systems, key-value stores, and information retrieval. Over the past 25 years, record-level locking has also become ubiquitous in database storage str...
-
Book
-
Chapter
Serializable Timestamp Validation
Optimistic concurrency control and its end-of-transaction validation logic may compare concurrent transactions’ read and write sets or it may focus on timestamps attached to database objects. The former techni...
-
Chapter
Avoiding Index-Navigation Deadlocks
Query execution plans in relational databases often search a secondary index and then fetch data from a table’s primary storage structure. In contrast, database updates usually modify first a table’s primary s...
-
Chapter
Deferred Lock Enforcement
Within a database transaction, the longest phases execute the application logic and harden the commit log record on stable storage. Two existing techniques, early lock release and controlled lock violation, el...
-
Chapter
Hierarchical Locking in B-Tree Indexes
Three designs of hierarchical locking suitable for B-tree indexes are explored in detail and their advantages and disadvantages compared.
-
Chapter
Controlled Lock Violation
In databases with a large buffer pool, a transaction may run in less time than it takes to log the transaction’s commit record on stable storage. Such cases motivate a technique called early lock release: imme...
-
Chapter
Orthogonal Key-Value Validation
In pessimistic concurrency control (i.e., locking), “repeatable count” transaction isolation (i.e., serializability) can be enforced efficiently and at a fine granularity. For example, orthogonal key-value loc...
-
Chapter
Repairing Optimistic Concurrency Control
Optimistic concurrency control means end-of-transaction validation of read- and write-sets or of timestamps attached to database items. Optimistic concurrency control suffers from multiple problems the present...
-
Reference Work Entry In depth
Parallel Query Optimization
-
Reference Work Entry In depth
Storage Manager
-
Reference Work Entry In depth
Buffer Pool
-
Reference Work Entry In depth
B-Tree Locking
-
Reference Work Entry In depth
Parallel Query Execution Algorithms
