Did anyone who read the magazine read the non-Informix articles?

IBM is clearly cognizant about the capabilities of Hadoop, but something is lost in translation. The DB2 centric authors don't understand
Hadoop/HBase so they can't communicate anything beyond the initial point...Hadoop really good at handling massive amounts of data.
Hadoop can output to your relational database as a sink.

IBM clearly also recognizes that what constitutes as a 'very large' database has lost its meaning.
An example is the National Mamogram database. They don't store the mamograms in the database, but just the indexing/meta data that describes the image. So is that really a large database?

And yet in these articles, IBM glosses over the reason why Hadoop/Hbase is attractive.

Using the mamo database, one could store the meta data information in HBase and have incredible fast access all at the price of a single license of IDS or DB2.

Is HBase perfect? No. Lots of gaps that you don't have in a mature seasonedRelational Database. (Relational Databases are 30+ years old. HBase? not so much.)

-G


__________________________________________________ _______________
Your E-mail and More On-the-Go. Get Windows Live Hotmail Free.
http://clk.atdmt.com/GBL/go/201469229/direct/01/