Because today we're wondering a whole lot about how to represent big data in memory over many machines if you're problem is in such a way divisable, or on one machine if possible because what we want to do is to draw conclusions from that data, by querying it, and if we find a way to both store big data cheaply in a form where it is still queryable, then we have solved a problem most businesses today tamper with, that is what do I do with all this data?
Businesses are getting crushed under the weight of their data. They "solve" it by stashing it onto Hadoop. Unfortunately, no one then knows how to map/reduce over that data. So they're none the wiser even though they kept the data that could have made their analysts smarter than their competitor's. This of course totally awsome if you're a Hadoop consultant.
Well, that's at least why I dwell on data structures.
Businesses are getting crushed under the weight of their data. They "solve" it by stashing it onto Hadoop. Unfortunately, no one then knows how to map/reduce over that data. So they're none the wiser even though they kept the data that could have made their analysts smarter than their competitor's. This of course totally awsome if you're a Hadoop consultant.
Well, that's at least why I dwell on data structures.
Edit: spelin