Performance of B+ tree concurrency control algorithms

Srinivasan, V.; Carey, Michael J.

doi:10.1007/bf01263046

Cited by 64 publications

(19 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Concurrency and recovery of B-link trees including node consolidations and deletions were studied in [11,12,13]. Blink trees were shown to have better concurrency than B-trees in [8,20].…”

Section: Lazy Merging Of B-link Treesmentioning

confidence: 98%

See 1 more Smart Citation

Online B-tree merging

Sun

Wang

Salzberg

et al. 2005

Proceedings of the 2005 ACM SIGMOD International Conference on Management of Data

View full text Add to dashboard Cite

Many scenarios involve merging of two B-tree indexes, both covering the same key range. Increasing demand for continuous availability and high performance requires that such merging be done online, with minimal interference to normal user transactions. In this paper we present an online B-tree merging method, in which the merging of leaf pages in two B-trees are piggybacked lazily with normal user transactions, thus making the merging I/O efficient and allowing user transactions to access only one index instead of both. The concurrency control mechanism is designed to interfere as little as possible with ongoing user transactions. Merging is made forward recoverable by following a conventional logging protocol, with a few extensions. Should a system failure occur, both indexes being merged can be recovered to a consistent state and no merging work is lost. Experiments and analysis show the I/O savings and the performance, and compare variations on the basic algorithm.

show abstract

“…Concurrency and recovery of B-link trees including node consolidations and deletions were studied in [11,12,13]. Blink trees were shown to have better concurrency than B-trees in [8,20].…”

Section: Lazy Merging Of B-link Treesmentioning

confidence: 98%

“…The smallest time unit in the simulation is one instruction time (one second 1G time units). The time cost of basic operations are listed inTable 3(partly based on[20] and communication with commercial DBMS implementors). A mergeEvent is not a database lock and it can be implemented by two semaphores.…”

mentioning

confidence: 99%

Online B-tree merging

Sun

Wang

Salzberg

et al. 2005

Proceedings of the 2005 ACM SIGMOD International Conference on Management of Data

View full text Add to dashboard Cite

show abstract

“…The basic structure of the B dH -tree is similar to that of the B link -tree [8], the latter of which is a version of the B + -tree that has been modified to support concurrency control techniques. Internal nodes on the same level of the B link -tree are linked for concurrency control algorithms.…”

Section: Naïve Approach Based On the Hilbert Curvementioning

confidence: 99%

BdH-Tree: A B+-Tree Based Indexing Method for Very Frequent Updates of Moving Objects

Seo

Song

Park

et al. 2008

International Symposium on Computer Science and Its Applications

View full text Add to dashboard Cite

Several methods to index positions of moving objects have been proposed. Most of them are based on R-tree-like index structures. Some studies have made efforts to improve the update performance of R-trees, and they actually focus on query performance. Even though the update performance has been improved by virtue of those researchers' efforts, the overhead and immaturity of the concurrency control algorithms of Rtrees make us hesitant to choose them for moving objects. In this paper, we propose an update-efficient indexing method. The proposed index is based on B + -tree and the Hilbert curve. We present an advanced Hilbert curve that automatically adjusts the order of the Hilbert curve in subregions, according to the data distribution. Through empirical study, we show that our strategy achieves a faster response time and higher throughput than those of other strategies.

show abstract

“…The B link -tree performs wait-free searches when implemented for the memory-disk boundary where a page can be accessed from disk atomically. When implemented in main memory, the tree uses shared reader-writer locks [21,22]. The reader-writer lock strategy is a bottleneck when there are only a handful of nodes in the data structure.…”

Section: Performancementioning

confidence: 99%

Lock-Free Multiway Search Trees

Spiegel

Reynolds

2010

2010 39th International Conference on Parallel Processing

View full text Add to dashboard Cite

We propose a lock-free multiway search tree algorithm for concurrent applications with large working set sizes. Our algorithm is a variation of the randomized skip tree. We relax the ordering constraints among the nodes in the original skip tree definition. Optimal paths through the tree are temporarily violated by mutation operations, and eventually restored using online node compaction. Experimental evidence shows that our lock-free skip tree outperforms a highly tuned concurrent skip list under workloads of various proportions of operations and working set sizes. The max throughput of our algorithm is on average 41% higher than the throughput of the skip list, and 129% higher on the workload of the largest working set size and read-dominated operations. I. IntroductionThe term 'memory wall' was coined over fifteen years ago to communicate the disparity between microprocessor speeds and memory latency and bandwidth in sequential processors [1]. Multicore architectures have increased the demand for low latency, high bandwidth memory access. A report on the landscape of parallel computing research identified the memory wall as a major obstacle to good performance for almost half of all classes of computational problems in the report [2]. A study of supercomputing applications at Sandia National Laboratories found that doubling memory latency leads to a 11% drop in the performance of floating-point applications and a 32% drop in the performance of integer applications [3].The lock-free skip list implements a concurrent ordered set abstract data type. The skip list is scalable under thread contention. However, the design exhibits poor spatial locality of reference. A traversal that accesses N elements in the list also access at least N nodes. A multiway search tree would improve spatial locality by storing several elements in a single node.In this paper we present a lock-free skip tree algorithm. Our algorithm implements a lock-free randomized multiway search tree. The lock-free skip tree has relaxed structural properties that allow atomic operations to modify the tree without invalidating the consistency of the data structure. Our skip-tree

show abstract

Performance of B+ tree concurrency control algorithms

Cited by 64 publications

References 29 publications

Online B-tree merging

Online B-tree merging

BdH-Tree: A B+-Tree Based Indexing Method for Very Frequent Updates of Moving Objects

Lock-Free Multiway Search Trees

Contact Info

Product

Resources

About