Created on 2014-08-08 12:30
Published on 2014-08-08 12:51
Conventional SQL Relational database management systems are essentially data silos by design -- due to their use of literals based identifiers in regards to primary keys, foreign keys, tables, columns, indexes, and stored procedures denotation (naming).
Unfortunately, the assertions above do not provide justification for "ripping and replacing" existing SQL Relational Database Management Systems (SQL RDBMS) installations due to the fact that they are now integral parts of enterprise IT infrastructure.
The issues outlined above present a conundrum due to the following contradictory facts:
[1] SQL RDBMS products are exponential vectors of enterprise inertia in regards to data exploitation -- each is a data silo with eternally increased data integration challenges
[2] SQL RDBMS products (even the NewSQL variety) are performance focused -- issues relating to identity, privacy, and security remain poor, due to data silo underpinnings
[3] Enterprise investment in SQL RDBMS products continues to grow at rates that still exceed those of alternatives
[4] Enterprises and Individuals alike continue to look to improved data access and integration as the keys to increasing agility and competitive advantage
In recent times, I've written blog posts covering data de-silo-fication in regards to Oracle [1], Microsoft SQL Server [1], and IBM DB2 [3] (the most recent) that include live examples of RDF based Linked Data views over ODBC accessible data sources.
Links:
[1] http://kidehen.blogspot.com/2014/04/oracle-data-de-silo-fication-using-rdf.html -- Oracle
[2] http://kidehen.blogspot.com/2014/06/sql-server-data-de-silo-fication-using.html -- Microsoft SQL Server
[3] http://kidehen.blogspot.com/2014/08/db2-sql-rdbms-data-de-silo-fication.html -- IBM DB2
[4] http://kidehen.blogspot.com/2014/10/progress-78910-andor-openedge-data-de.html -- Progress (OpenEdge) .