Data Models and Query Languages for Data Streams
Author
Søndergaard, Jes
Term
4. term
Education
Publication year
2005
Abstract
Dette studie undersøger, hvad der gør datastreams forskellige fra traditionelle relationsdatabaser. En datastream er et kontinuerligt, potentielt uendeligt flow af data, i modsætning til de endelige tabeller (“relationer”) i relationsdatabasesystemer. Forskere er uenige om, hvorvidt streams er et nyt dataparadigme eller blot en udvidelse af relationer. For at belyse dette gennemgår studiet datamodeller og forespørgselssprog i eksisterende stream-systemer og sammenligner dem med dem for relationsdata. Undersøgelsen finder, at streams har egenskaber, der adskiller dem, især deres kontinuerlige og ubegrænsede natur. Alligevel er forskellene så begrænsede, at stream-behandling kan implementeres oven på eksisterende databasesystemer.
This study explores what makes data streams distinct from traditional relational databases. A data stream is a continuous, potentially unending flow of data, unlike the finite tables (“relations”) stored in relational database systems. Researchers disagree on whether streams are a new data paradigm or simply an extension of relations. To investigate, the study reviews the data models and query languages used in existing stream systems and compares them with those for relational data. The findings show that streams have characteristics that set them apart, especially their continuous and unbounded nature. Even so, the differences are limited enough that stream processing can be implemented on top of existing database systems.
[This abstract was generated with the help of AI]
Documents