Re:Billions wasted on Hadoop startups, the same will eventua...
by vkazanov on Saturday June 01, 2019 @03:11AM
(#20067788)
The thing is that this is exactly my company's scale, and hdfs is perfectly fine here. :-)
And I am curious about possible alternatives: open source, about 100-200 machines, with good support for analytics and SQlish systems.
BTW, "hundreds of terabytes on hundreds of machines" isn't interesting territory any more. Most people's needs are far smaller, so HDFS isn't much help. Those who need more generally need much more, so HDFS isn't much help again. Richer semantics are nice either way. Imagine thousands of machines with dozens of terabytes each and you might start to see the problems with HDFS's design (though you'll still be far short of the domain I work in).