IBM®
Skip to main content
    United States [change]    Terms of use
 
 
 
    Home    Products    Services & solutions    Support & downloads    My account    
IBM Research

(none)

Impliance -- An Information Management Appliance


Overview

The goal of the Impliance project is to build a next-generation information management system that stores all structured, unstructured, and semi-structured data, is easy to manage, and analyzes data in a scalable way. The long-term vision of Impliance has been articulated in our 2007 CIDR paper (2007CIDRvisionPaper). In 2007, the goal of Impliance is to build a warehouse infrastructure that stores large amounts of data with text, extracts annotations through large-scale text analytics, and exploits annotations for semantic search and BI analysis. This infrastructure will be scalable, fault-tolerant, and easy to manage.

Technical Goals

  • Build a scalable data store and computing infrastructure that scales down to a 1 TB desktop and scales up to thousands of nodes and many thousands of TBs of data.
  • Build a flexible data store that can efficiently store, retrieve, and index data consisting of many different types of data, ranging from small relational tuples to short e-mail messages, huge video files, and annotations on all those data types.
  • Build a usable computing infrastructure that allows computations to be composed of existing operations, allows users with special needs to write their own custom operators, and supports both batch analysis over millions of documents and thousands of queries per second over indexed documents.
  • Build a management infrastructure that allows the device to be self managing.

Members

  • Project Leads
    • Guy Lohman
    • Eugene Shekita 
  • Team Members
    • Kevin Beyer
    • Vuk Ercegovac
    • Ning Li
    • Mauricio Mediano
    • Hamid Pirahesh
    • Jun Rao
    • Fred Reiss

External Collaborator

  • To the maximum extent possible, we are incorporating open source components into Impliance, when they meet our objectives. For example, we are building Impliance on top of the Hadoop open source distributed file system, and are very likely to use Lucene for indexing.

Selected Publications

  • Bishwaranjan Bhattacharjee, Vuk Ercegovac, Joseph Glider, Richard Golding, Guy Lohman, Volker Markl, Hamid Pirahesh, Jun Rao, Robert Rees, Frederick Reiss, Eugene Shekita, Garret Swart, "Impliance: A Next Generation Information Management Appliance", Procs. of the Third Biennial Conference on Innovative Data Systems Research (CIDR 2007), Monterey, CA, Jan. 2007, pp. 351-360. 2007CIDRvisionPaper

   Overview

Members

External Collaborators

Selected Publications


    About IBMPrivacyContact