Linked Data In The Enteprise

1,867 views
1,788 views

Published on

ESTC 2009 panel

Linked Data in the Enterprise - is it just another hype or does it have real added value for information management?

Published in: Technology, Education
0 Comments
3 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,867
On SlideShare
0
From Embeds
0
Number of Embeds
5
Actions
Shares
0
Downloads
0
Comments
0
Likes
3
Embeds 0
No embeds

No notes for slide

Linked Data In The Enteprise

  1. 1. ESTC 2009 panel: Linked Data in the Enterprise - is it just another hype or does it have real added value for information management? Panel statement of Atanas Kiryakov Dec, 2009
  2. 2. Introduction <ul><li>Plenty of people in the IT world have heard about linked data and like it </li></ul><ul><li>However, linked data still have limited impact in the enterprises </li></ul><ul><li>Because: </li></ul><ul><ul><li>There is no well established opinion and practices what linked data can be “buy” for the enterprise? </li></ul></ul><ul><ul><ul><li>What are the concrete benefits? </li></ul></ul></ul><ul><ul><li>It is not clear what would it cost? </li></ul></ul><ul><ul><ul><li>What are the associated risks? </li></ul></ul></ul><ul><li>I will present our answers to these questions </li></ul>ESTC Panel: Linked Data in the Enterprise Dec, 2009
  3. 3. Linked Data in the Enterprise: Challenges <ul><li>LOD is hard to comprehend </li></ul><ul><ul><li>Diversity comes at a price </li></ul></ul><ul><ul><li>One needs to make a query against 200 different schemata and hundreds of thousands of classes and properties </li></ul></ul><ul><li>LOD is unreliable </li></ul><ul><ul><li>Most of the servers behind LOD today are slow </li></ul></ul><ul><ul><ul><li>High down time </li></ul></ul></ul><ul><ul><li>Dealing with data distributed on the web is slow </li></ul></ul><ul><ul><ul><li>A federated SPARQL query, which uses 2-3 servers within several joins can be *very* slow </li></ul></ul></ul><ul><ul><li>No sort of consistency is guaranteed </li></ul></ul><ul><ul><ul><li>Low commitment to the formal semantics and intended usage of the ontologies and schemata </li></ul></ul></ul>ESTC Panel: Linked Data in the Enterprise Dec, 2009
  4. 4. Linked Data in the Enterprise: Why? <ul><li>To facilitate data integration </li></ul><ul><ul><li>One can use LOD as “interlingua” for enterprise data integration </li></ul></ul><ul><ul><li>Additional public information can help alignment and linking </li></ul></ul><ul><li>To add value to proprietary data </li></ul><ul><ul><li>Public data can allows more analytics on top of proprietary data </li></ul></ul><ul><ul><li>For instance, by linking to spatial data from Geonames </li></ul></ul><ul><li>Make enterprise data more open </li></ul><ul><ul><li>To make the enterprise data easier to use outside the enterprise </li></ul></ul><ul><ul><li>Public identifiers and vocabulary can be used to access them </li></ul></ul>ESTC Panel: Linked Data in the Enterprise Dec, 2009
  5. 5. Linked Data in the Enterprise: Challenges <ul><li>LOD is hard to comprehend </li></ul><ul><ul><li>Diversity comes at a price </li></ul></ul><ul><ul><li>One needs to make a query against 200 different schemata and hundreds of thousands of classes and properties </li></ul></ul><ul><li>LOD is unreliable </li></ul><ul><ul><li>Most of the servers behind LOD today are slow </li></ul></ul><ul><ul><ul><li>High down time </li></ul></ul></ul><ul><ul><li>Dealing with data distributed on the web is slow </li></ul></ul><ul><ul><ul><li>A federated SPARQL query, which uses 2-3 servers within several joins can be *very* slow </li></ul></ul></ul><ul><ul><li>No sort of consistency is guaranteed </li></ul></ul><ul><ul><ul><li>Low commitment to the formal semantics and intended usage of the ontologies and schemata </li></ul></ul></ul>ESTC Panel: Linked Data in the Enterprise Dec, 2009
  6. 6. Reason-able views to the LOD <ul><li>Reason-able views represent an approach for reasoning and management of linked data </li></ul><ul><li>Key ideas: </li></ul><ul><ul><li>Group selected datasets and ontologies in a compound dataset </li></ul></ul><ul><ul><li>Load the compound dataset in a single semantic repository </li></ul></ul><ul><ul><li>Perform inference with respect to tractable OWL dialects </li></ul></ul><ul><li>Objectives </li></ul><ul><ul><li>Make reasoning and query evaluation feasible </li></ul></ul><ul><ul><li>Guarantee a basic level of consistency </li></ul></ul><ul><ul><li>Guarantee availability </li></ul></ul><ul><ul><li>Easier exploration and querying of unseen data </li></ul></ul><ul><ul><ul><li>Lower the cost of entry through URI auto-complete and RDF search </li></ul></ul></ul>ESTC Panel: Linked Data in the Enterprise Dec, 2009
  7. 7. Linking Open Data Datasets and Views (red and yellow) Dec, 2009 ESTC Panel: Linked Data in the Enterprise LOD map from: http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData

×