From: wizofoz2k@yahoo.com.au   
      
   On 12/05/2016 1:14 @wiz, Mladen Gogala wrote:   
      
   >> Like you, I wish Tanel all the best with his new company.  Not sure what   
   >> it does but likely not very important to my circles.   
   >   
   > Actually, it does look  interesting:   
   > ************************************************************************   
   > Offload   
   > Gluent Offload Engine automates offloading data from enterprise databases   
   > to Hadoop. You can have an up-to-date copy of your data, in a familiar   
   > data model, ready for analytics in the powerful ecosystem of Hadoop.   
   > ************************************************************************   
      
   Interesting!  Thanks for that, good to know. Looks like they use a redo   
   log read mechanism along the same lines as Delphix and likely inspired   
   by Goldengate. Only way to ensure synch without internals and shared   
   memory peeking.   
      
      
   > Hadoop is just one of the "Big Data" technologies spawned by exorbitant   
   > prices of Oracle RDBMS.   
      
   We been looking at the whole Hortonworks stack and it looks very good.   
   But the problem with all these BigData products is always the same: they   
   require a fully dedicated person who ends up being the only one who can   
   mine and extract data.  Mostly because of the absence of any schema   
   metadata. While with relational dbs anyone with basic knowledge of SQL   
   can pop-down and start producing consistent reports and logic in a very   
   short time.   
      
   Yes, we're back to the bad old days of hierarchical and network db data   
   lockdown...   
      
      
   > I have recently participated in a rather interesting project of replacing   
   > CLOB entries in Oracle by documents in MongoDB. The trick was to make the   
   > stuff searchable and there is an excellent open source search engine   
   > called Sphinx, that I've used. Basically, the CLOB in Oracle started   
   > consuming a huge amount of database space and Oracle suggested advanced   
   > compression and deduplication to deal with that. MongoDB was much cheaper   
   > option and Sphinx can do text search with the best of them. Web interface   
   > was written in PHP.   
      
   Cool! Thanks for sharing that - it's quite relevant to us as well. CLOBs   
   - and LOBs in general - are the Achilles Heel of Oracle.  High time   
   Oracle development got the clear message to fix the blessed things and   
   make *LOBs truly usable instead of a source of easy consultant contract   
   work to clean-up/re-organize/control/manage.   
   But unfortunately - and this is IMHO one of the abysmal failures of the   
   entire ACE program - they totally ignored user and ACE feedback on   
   possible improvements and directions for the product.  I know from   
   talking with various ACEDs they asked for a lot of improvements over the   
   years which never got the light of day. Most unfortunate.   
      
   Ah well, I'm quite sure Larry's announcement on the 1st June is going to   
   change the entire cloud.   
   Or something...  :P   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)