Hacker News new | past | comments | ask | show | jobs | submit login
IBM Unveils Software Defined Storage Technology for Era of Big Data (ibm.com)
27 points by cjdulberger on May 13, 2014 | hide | past | favorite | 6 comments



Elastic Storage is the product formerly known as GPFS. http://www-03.ibm.com/systems/technicalcomputing/platformcom...


...and GPFS stands for "General Parallel File System."


Me, I prefer Lustre over GPFS; Lustre is open source as well, and used by 70% of Top100

Here's a related presso from KIT:

https://www.scc.kit.edu/scc/docs/Lustre/pfs_scc-sem_20111005...


So IBM sells UIMA with Lucene on top of GPFS and calls it Software Defined Storage.

* UIMA (Unstructured Information Management Architecture): http://en.wikipedia.org/wiki/Unstructured_Information_Manage...

* Lucene: http://en.wikipedia.org/wiki/Lucene

* GPFS: http://en.wikipedia.org/wiki/GPFS

Edit: GPFS instead of Hadoop


Nope, that's not what this product is at all. It's GPFS. You could bang all three of those together fifty ways and not reinvent a distributed filesystem, so I have no idea what you're talking about. (Seriously? Lucene?)

Speaking as (I'm sure) one of the few people in this thread who will have used GPFS: GPFS is actually quite good. You'd be surprised how good it is. It has its downfalls, but it is extraordinarily useful, performs well, and just about hits the holy grail of distributed POSIX. It's also absurdly expensive, but: IBM.


Has anybody felt that IBM is absurdly expensive? Why should I pay tons of money to use a proprietary filesystem type?




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: