In-memory processing: Difference between revisions

Revision as of 19:07, 17 October 2012

Revision as of 19:23, 17 October 2012

Line 46:

== In-memory market Vendors ==

The idea of running memory based databases was first developed by [[QlikTech]] in 1997 with their business intelligence product QlikView. Since lower costs are one of the benefits of Inin-memory processing many organizations are looking to adopt this technology and many vendors have since added in-memory to their platforms. Biggies like SAP recently unveiled High-Performance Analytical Appliance (HANA) for in-memory computing, Oracle acquired [[TimesTen]], an in-memory relational database. IBM Cognos (formerly Applix TM1) offers financial application and have many financial institutions as customers. Products such as Spotfire acquired by TIBCO, IBM SolidDB are already popular and have made their mark.<ref>{{cite journal|last=Henschen|first=Doug|title=Next-Gen BI Is Here|journal=Information Week|date=31|year=2009|month=August|issue=1239|pages=6|url=http://www.businessintelligence.info/docs/revistas/bispain_tendencias_business_intelligence.pdf}}</ref>

Another solution, SiSense's Prism software, uses elastic in-memory processing combined with a columnar data store (see [[Column oriented DBMS]]). The column-oriented database allows for faster queries; joins are completed in RAM using algebraic arguments, as opposed to the traditional table join of an [[RDBMS]]. Only a small fraction of a data set is held in RAM at any given moment.<ref>{{cite web|last=Israeli|first=Elad|title=In-Memory BI is not the Future|publisher=The Elasticube Chronicles|pages=6|url=http://elasticube.blogspot.com/2010/09/in-memory-bi-is-not-future-its-past.html}}</ref>

== References ==