So recently I posted some thoughts on big data and the increasing usage of Hadoop, the general theme was data management != data analysis…this caused confusion with some folks, as evidenced by the twitter exchange (tweets haven’t been altered but some extraneous ‘noise’ removed to maximize your reading pleasure)
@Beaker @amrittsering I’m confused by your last blog. Is your point that people are spending $$$ on data aggregation hoping it leads to analytics?
I read/re-read your posts & it’s almost like u r suggesting majority of co’s deploying Hadoop (e.g) are clueless WRT why?
@amrittsering @Beaker it isn’t that they’re clueless, data mngt is important, but without the ability to analyse managing data isn’t what is being claimed
@Beaker @amrittsering don’t disagree but I don’t see co’s pouring millions into hadoop “hoping” sthing pops out. They have analytic sw w/map reduce
@amrittsering @beaker From an analytics perspective the challenge with Hadoop is it shifts the burden from ingesting and processing data to querying data
@amrittsering @Beaker Having to map->reduce isn’t easy and developing queries is challenging even with hive, pig and other tools…
@amrittsering @Beaker Hadoop hasn’t magically removed all the problems of data warehousing, managing 1,000’s of nodes, map-reducing, HDFSing, etc…!=easy
@amrittsering @Beaker The problem is that many analytics tools have been designed for RDBMS, column-store, OLAP cubes, or other approaches – more problems
@amrittsering @Beaker & new tools will emerge to add powerful analytics on top of new data storage approaches, & as my post says the marktet will evolve
@amrittsering Beaker So to recap, hadoop !=analytics, analytics tools need to evolve to provide value against new approaches, in the meantime the trough
@Beaker @amrittsering If it’s such a simple issue, why’s it taking 4 blog posts to say? < that was the point of confusion. U said it in < 140 chars!
Because I suck at blogging…and twitter