Matches in SemOpenAlex for { <https://semopenalex.org/work/W2490876026> ?p ?o ?g. }
Showing items 1 to 55 of
55
with 100 items per page.
- W2490876026 abstract "In this dissertation, an array processor connected in an n-dimensional mesh is evaluated for very large relational database applications. Each node is assumed to have substantial processing capability with substantial amount of memory. By partitioning each base relation among all nodes, and by storing them directly within primary memory of a memory-resident database machine, the bottleneck of slow secondary storage access is eliminated. Whenever inter-node operations are required, local partitions are routed to other nodes.In this dissertation, an n-dimensional mesh is studied for SIMD database processing. Database algorithms are classified into five categories based on their routing requirements: (1) database algorithms which do not require any routing such as select; (2) database algorithms that require partial routing without order such as tuple balancing; (3) database algorithms that require partial routing with order such as project, elimination of duplicates, and sort; (4) database algorithms that require full routing such as nested-loop, sort-merge join algorithms, and cartesian product; (5) database algorithms that require random routing such as hash-based join and hash-based aggregate algorithms. Routing algorithms appropriate for these different classes of database algorithms are presented, and their performances are analyzed. Their performances on an n-dimensional mesh are compared with their performances on a binary cube, which is a subset of the n-dimensional mesh. Incremental expansion is studied on an n-dimensional mesh and its impact on performance is measured.The design of a large parallel computer is often limited by the longest interconnection. In an n-dimensional mesh, the method of interconnection, individual interconnection length and node architecture remain identical as the computer is expanded. A custom node architecture maximizes processing and communication parallelism. Another node architecture uses standard microprocessor components.Several n-dimensional meshes are connected through serial communication links in the (n + 1)th dimension to support even larger database applications. Each n-dimensional mesh may execute a different database operation or query resulting in an MIMD database computer. Two query execution strategies are presented and their performance are studied using simulation models." @default.
- W2490876026 created "2016-08-23" @default.
- W2490876026 creator A5036851419 @default.
- W2490876026 date "1992-01-01" @default.
- W2490876026 modified "2023-09-23" @default.
- W2490876026 title "Database processing on an n-dimensional mesh" @default.
- W2490876026 hasPublicationYear "1992" @default.
- W2490876026 type Work @default.
- W2490876026 sameAs 2490876026 @default.
- W2490876026 citedByCount "0" @default.
- W2490876026 crossrefType "journal-article" @default.
- W2490876026 hasAuthorship W2490876026A5036851419 @default.
- W2490876026 hasConcept C11413529 @default.
- W2490876026 hasConcept C173608175 @default.
- W2490876026 hasConcept C188805328 @default.
- W2490876026 hasConcept C2777541480 @default.
- W2490876026 hasConcept C38652104 @default.
- W2490876026 hasConcept C41008148 @default.
- W2490876026 hasConcept C77088390 @default.
- W2490876026 hasConcept C99138194 @default.
- W2490876026 hasConceptScore W2490876026C11413529 @default.
- W2490876026 hasConceptScore W2490876026C173608175 @default.
- W2490876026 hasConceptScore W2490876026C188805328 @default.
- W2490876026 hasConceptScore W2490876026C2777541480 @default.
- W2490876026 hasConceptScore W2490876026C38652104 @default.
- W2490876026 hasConceptScore W2490876026C41008148 @default.
- W2490876026 hasConceptScore W2490876026C77088390 @default.
- W2490876026 hasConceptScore W2490876026C99138194 @default.
- W2490876026 hasLocation W24908760261 @default.
- W2490876026 hasOpenAccess W2490876026 @default.
- W2490876026 hasPrimaryLocation W24908760261 @default.
- W2490876026 hasRelatedWork W1964497035 @default.
- W2490876026 hasRelatedWork W2003541146 @default.
- W2490876026 hasRelatedWork W201013271 @default.
- W2490876026 hasRelatedWork W2022106793 @default.
- W2490876026 hasRelatedWork W2076005690 @default.
- W2490876026 hasRelatedWork W2097692226 @default.
- W2490876026 hasRelatedWork W2130801009 @default.
- W2490876026 hasRelatedWork W2132727211 @default.
- W2490876026 hasRelatedWork W2140869148 @default.
- W2490876026 hasRelatedWork W2165876198 @default.
- W2490876026 hasRelatedWork W2292315185 @default.
- W2490876026 hasRelatedWork W2347597937 @default.
- W2490876026 hasRelatedWork W2378725315 @default.
- W2490876026 hasRelatedWork W2395019391 @default.
- W2490876026 hasRelatedWork W2511292067 @default.
- W2490876026 hasRelatedWork W2529787296 @default.
- W2490876026 hasRelatedWork W2540049111 @default.
- W2490876026 hasRelatedWork W2901210942 @default.
- W2490876026 hasRelatedWork W2963815030 @default.
- W2490876026 hasRelatedWork W2979858238 @default.
- W2490876026 isParatext "false" @default.
- W2490876026 isRetracted "false" @default.
- W2490876026 magId "2490876026" @default.
- W2490876026 workType "article" @default.